agent-github-modes
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- Prompt Injection (LOW): Detected indirect prompt injection surface (Category 8). \n
- Ingestion points: The skill ingests untrusted data from external sources using
gh pr view,gh issue list, and theReadtool (File: SKILL.md).\n - Boundary markers: There are no explicit delimiters or instructions provided to the agent to distinguish between internal instructions and external data content.\n
- Capability inventory: The skill possesses the
Bashtool,Writecapability, and agent spawning tools, creating a path for malicious instructions in PRs or issues to be executed (File: SKILL.md).\n - Sanitization: No mechanisms for sanitizing or filtering external GitHub content are implemented.\n- Command Execution (SAFE): The skill utilizes the
Bashtool to perform repository operations such asgitandghCLI commands. This is consistent with the primary purpose of managing GitHub workflows.
Audit Metadata