code-review
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted code review feedback (references/code-review-reception.md), creating a surface for potential injection. However, it incorporates a mandatory response pattern (READ -> UNDERSTAND -> VERIFY) that serves as a logical boundary. The skill's capabilities include git commands, test/build command execution, and subagent dispatch via the Task tool. Sanitization is achieved through technical verification against the codebase using grep and a requirement for evidence-based verification before any claims are made.
- [COMMAND_EXECUTION]: The skill uses git utilities (git rev-parse, git log) and executes verification commands such as tests and builds. These are standard operations for a development-focused agent and are performed within a structured protocol that requires reading and verifying the command output.
- [SAFE]: The skill's instructions emphasize technical correctness and evidence over performative agreement, promoting secure and reliable agent behavior.
Audit Metadata