update-pr
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill requires the agent to analyze external, untrusted data (GitHub PR reviews and comments) to determine the development state and decide whether to perform destructive history rewrites (git rebase/force push).
- Ingestion points: Pull Request review status and comment history, accessed via
gh(GitHub CLI) commands (SKILL.md). - Boundary markers: The instructions tell the agent to "ignore tech noise" and distinguish between user, AI, and human comments, but do not provide strict delimiters or instructions to treat the comment text as inert data.
- Capability inventory: The skill authorizes high-impact shell commands including
git commit --fixup,git rebase --autosquash,git push --force-with-lease, andgh stackoperations (SKILL.md). - Sanitization: There is no evidence of filtering or sanitization of the comment metadata or body text before it is evaluated by the agent.
Audit Metadata