dev-code-reviewer
Pass
Audited by Gen Agent Trust Hub on Jul 2, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted code changes and pull request descriptions, creating a surface for indirect prompt injection.
- Ingestion points: The skill is triggered by commands to review code, PRs, or diffs, necessitating the ingestion of external source code and commentary (SKILL.md).
- Capabilities: The instructions involve high-level code reasoning and directives to execute various CLI tools for validation, such as Semgrep, tsc, and pip-audit (SKILL.md).
- Boundary markers: The 'Review Posture' section (REVIEW-POSTURE-01) explicitly directs the agent to treat external claims, AI summaries, and developer prose as untrusted until independently verified (SKILL.md).
- Sanitization: The skill does not define technical sanitization, escaping, or schema validation for the ingested code changes.
Audit Metadata