code-review-expert
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data in the form of
git diffoutput, which could theoretically contain malicious instructions embedded in code comments or strings intended to influence the agent's analysis. - Ingestion points: Code changes are read via
git diffandgit status(SKILL.md). - Boundary markers: The skill does not use specific delimiters for the diff content, but it uses a structured output format for its own findings.
- Capability inventory: Uses
git,rg, andgrepfor reading; potential file writes are protected by a confirmation step (SKILL.md). - Sanitization: The workflow explicitly requires user confirmation before any implementation ('Fix all', 'Fix P0/P1 only'), providing a critical human-in-the-loop safeguard.
- [SAFE]: The skill uses standard local development tools and does not perform unauthorized network operations, exfiltrate data, or execute remote code.
Audit Metadata