receiving-code-review
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill explicitly defines a framework for processing external, untrusted data (code review feedback). It mitigates injection risks by instructing the agent to verify requirements against the codebase reality, ask clarifying questions, and push back against technically unsound suggestions rather than implementing them blindly.
- [COMMAND_EXECUTION]: The instructions reference standard command-line tools such as
grepfor searching the codebase and the GitHub CLI (gh api) for managing pull request comment replies. These operations are restricted to the local environment and the project's official repository context.
Audit Metadata