已审阅并标注「工具集成」的研究与资产。
条目依据已审阅标签归类，同一主题不表示资源之间存在依赖关系。日期保留来源精度，注释保留原语言。
UltraCUA 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.
Introduces 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.
GUI-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.
WALT 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.
Ferret-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.
Introduces 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.
MAS-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.
UI-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.
Introduces 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.
Combines API and GUI actions with distributed desktop reinforcement learning and alternating supervised training to address training instability.
CoAct-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.
Magentic-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.
MCPWorld 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.
AIOS 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.
Combines 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.
Introduces 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.
SkillWeaver 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.
Compares API-driven and GUI-driven agents across architecture, development, and interaction considerations, discussing when hybrid approaches are useful.
Introduces a modular visual web-agent stack with planning and memory components, exposed through a remote-browser application and Chrome extension.
Introduces a visual IDE environment and benchmark for software-engineering tasks, comparing screenshot-based control with access to editing and shell tools.
Recasts mobile UI automation as on-device code generation, using app-specific documentation and synthetic tasks to guide smaller models.
Introduces data and a generation pipeline for translating natural-language requests into Android intent invocations using compact language models.
Introduces an orchestrated team of browsing, file, and coding agents with progress tracking and error recovery for general digital tasks.
OSCAR 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.
Introduces a platform that integrates specialized agents under a coordinating MetaAgent for varied computer tasks.
Compares browsing-only, API-only, and hybrid web agents on WebArena, examining how access to application APIs changes task execution alongside browser interaction.
Introduces 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.
Introduces simulated application APIs and state-based task evaluation for interactive coding agents performing everyday digital workflows.
Introduces 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.
Evaluates mobile agents using combined UI and API actions, tasks with varying app scope, and checkpoints for intermediate progress.
Introduces 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.
Provides an operating-system agent framework connecting applications, files, terminals, and web tools, with accumulated skills for subsequent tasks.
Introduces questions that test assistant reasoning, multimodal understanding, browsing, and tool use beyond screen-action execution alone.
Evaluates multi-turn PowerPoint creation and editing, scoring the resulting presentation files rather than requiring a fixed API sequence.
WebWISE 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.
OpenAgents 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.
Introduces SheetCopilot, an agent that steers spreadsheet software via atomic virtual APIs, along with a 221-task spreadsheet manipulation benchmark.
Agent evaluation suite with containerized tasks and a function-calling interface, including operating-system and shopping environments.
Simulated applications, user data, and state-based task checks for evaluating agents that interact through APIs and code.
Browser-rendered mobile simulation with resettable app state and a shared task suite for GUI-only and GUI-plus-tool agents.
A containerized Android evaluation environment with GUI-only, agent-user interaction and MCP-augmented tasks, plus checks on application and backend state.
PowerPoint creation and editing benchmark with API-based execution and checks of the resulting presentation files.
An 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.
Instruction-to-function-call examples for learning Android intent invocation, with training, testing and API-annotation files for mobile task execution.
Mobile-agent supervision from emulator and simulator trajectories, plus executable tasks with starting states and success checks for reinforcement learning.
A Qwen3.6-based computer-use checkpoint designed to request GUI actions, code execution and tool calls through an external agent harness.
A released MAI-UI checkpoint combining mobile GUI actions, agent-user interaction and tool-use-oriented navigation.
Mobile-agent checkpoint based on Qwen3.6-27B for GUI and app-native tool interaction.
Mobile-agent checkpoint trained on OpenMobile-Data-v2 for screenshot actions and app-native tool calls.
A browser-automation command-line interface that exposes browser state and interaction commands for AI agents.
An Android automation implementation that builds app documentation from exploration traces and evaluates language-model-generated execution code in emulator workflows.
A Gradio interface for configuring and running Browser Use agents, including persistent browser sessions and interaction recordings.
Chromium-based browser implementations for external agent control and built-in assistant workflows, with session visibility and replay facilities.
Agent framework for browser and desktop tasks using automation code.
Data-generation, fine-tuning and evaluation utilities for teaching small language models to invoke Android intents through predefined functions.
Modular web-agent framework with browser control, planning, workflow memory, and tree-search components, plus web and extension interfaces.
第 61–120 条，共 136 条资源。
通过主题导出清单获取完整记录和证据。克隆目录后，可将主题记录直接交给标准 RAG 与引用导出工具。
