computer-use

Pass

Audited by Gen Agent Trust Hub on Aug 19, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute local CLI binaries (such as orca, orca-ide, or orca-dev) to inspect and control desktop application windows. It includes specific logic for resolving the executable path based on environment variables like ORCA_CLI_COMMAND and ORCA_DEV_REPO_ROOT.
  • [INDIRECT_PROMPT_INJECTION]: The skill allows the agent to ingest data from external sources (desktop application UI, accessibility trees, and screenshots) which can contain adversarial instructions.
  • Ingestion points: Data is ingested from any local desktop application window or browser webview via the orca computer-use CLI commands.
  • Boundary markers: The skill does not define specific delimiters or instructions to prevent the agent from obeying commands embedded within the UI it is inspecting.
  • Capability inventory: The skill grants the agent extensive control over the local system, including the ability to type, click, scroll, and set values within other applications.
  • Sanitization: No sanitization, filtering, or validation of the UI content is performed before it is presented to the agent's context.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 19, 2026, 06:52 AM
Security Audit — agent-trust-hub — computer-use