computer-use

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Uses the okou computer-use CLI to list, open, and inspect applications, as well as perform UI actions like clicking, typing, and scrolling. These commands allow the agent to directly interact with the host's graphical user interface.
  • [DYNAMIC_EXECUTION]: Utilizes node -e commands to parse JSON output and filter accessibility trees. This dynamic execution is used for data processing of local CLI outputs generated by the toolset.
  • [INDIRECT_PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection by reading accessibility trees from third-party applications.
  • Ingestion points: Reads appState metadata and content from host applications via okou computer-use get-app-state (SKILL.md).
  • Boundary markers: Absent; the skill does not use specific delimiters or instructions to ignore commands found within application data.
  • Capability inventory: Desktop control capabilities including click, set-value, type-text, and press-key across the host system.
  • Sanitization: No sanitization is performed on the ingested application state before the agent processes it for decision-making.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 01:23 AM
Security Audit — agent-trust-hub — computer-use