code-review
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [Indirect Prompt Injection] (SAFE): The skill is designed to process external code review feedback, which constitutes an attack surface. However, it mitigates this risk through a mandatory 'VERIFY' protocol. (1) Ingestion: feedback from external reviewers in code-review-reception.md. (2) Boundaries: Explicit response patterns and verification gates. (3) Capabilities: Git, grep, and subagent tasking. (4) Sanitization: Mandatory verification against test/build results and codebase reality.
- [Command Execution] (SAFE): Standard git commands (rev-parse, log) and grep are used for legitimate development workflow tasks like commit tracking and code search. No high-risk or arbitrary command execution was detected.
- [Prompt Injection] (SAFE): Behavioral instructions such as 'no performative agreement' and the 'Circle K' signal are used to enforce technical rigor and persona consistency; they do not target the agent's safety filters or attempt to override system constraints.
Audit Metadata