maintain
Pass
Audited by Gen Agent Trust Hub on Apr 20, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill enforces a strict security policy requiring the agent to show the exact text of any proposed comment or action to the maintainer for explicit approval before execution. This 'human-in-the-loop' requirement effectively mitigates risks associated with automated repository management.- [COMMAND_EXECUTION]: Shell commands are used for legitimate project maintenance tasks, such as searching issues and PRs via the GitHub CLI ('gh') and inspecting the local codebase using Git. These commands are scoped to the specific project and do not perform unauthorized operations.- [PROMPT_INJECTION]: There is an inherent risk of indirect prompt injection as the skill processes untrusted data from GitHub issue threads. \n
- Ingestion points: Reads the full thread for every issue, including comments and body (Phase 2). \n
- Boundary markers: Not utilized during the data ingestion phase. \n
- Capability inventory: GitHub CLI ('gh') for issue/PR management and local file system search (Phases 0, 3, 4, and 6). \n
- Sanitization: No explicit sanitization of external content is performed. \n This risk is well-mitigated by the requirement for a human to review and approve all drafted responses before they are posted.
Audit Metadata