issuekit
Pass
Audited by Gen Agent Trust Hub on Oct 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill operates as a CLI wrapper for GitHub and Git management without performing unauthorized network requests or executing remote code. It relies on local authenticated tools (gh, git) and implements clear user confirmation steps for all impactful operations.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from GitHub issues and pull request descriptions.\n
- Ingestion points: Data enters the context via
gh issue view,gh pr list, and local plan files.\n - Boundary markers: The skill uses specific functional markers (e.g.,
Closes #N,Grilled: YYYY-MM-DD) to parse content, though it lacks explicit instructions to ignore adversarial content within these fields.\n - Capability inventory: Uses
ghCLI for tracker mutations andgitfor filesystem worktree management.\n - Sanitization: Risk is effectively mitigated by the mandatory human-in-the-loop requirement where all mutations (except labels) must be previewed and approved by the user before execution.
Audit Metadata