{
  "aliases": [],
  "authors": [],
  "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"
    }
  ],
  "id": "github:wuklab/osworld-human",
  "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",
  "notes": "The 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.",
  "published_at": null,
  "relations": [
    {
      "evidence_url": "https://raw.githubusercontent.com/WukLab/osworld-human/HEAD/README.md",
      "target_id": "arxiv:2506.16042",
      "type": "documented_by"
    }
  ],
  "review_status": "curated",
  "scope": "core",
  "summary": "Human reference trajectories and scoring code for measuring the action efficiency of computer-use agents on OSWorld tasks.",
  "topics": [
    "evaluation",
    "efficiency",
    "workflow"
  ],
  "updated_at": null,
  "url": "https://github.com/WukLab/osworld-human"
}