receiving-code-review
Pass
Audited by Gen Agent Trust Hub on Apr 11, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to use
gh apifor replying to GitHub pull request comments andgrepfor searching the codebase. These are standard operations for a development-focused agent and are used here within their intended context. - [PROMPT_INJECTION]: The instructions mandate a specific behavioral shift, requiring the agent to avoid common polite phrases (e.g., 'You're absolutely right!') and instead focus on technical verification. While this overrides default persona behaviors, it is aligned with the skill's stated purpose of ensuring technical correctness and does not attempt to bypass safety filters.
- [SAFE]: The skill inherently addresses the risk of indirect prompt injection by establishing a mandatory 'Verify before implementing' workflow. By requiring the agent to evaluate external feedback against the codebase reality and YAGNI principles before taking action, it provides a functional defense against malicious or incorrect suggestions embedded in external reviews.
Audit Metadata