review-agent
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues were identified. The skill follows best practices by instructing the agent to treat external content (diffs) as untrusted and explicitly forbids file modifications, commits, or network activity during the review process.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted code diffs and repository content, which is a standard requirement for its primary function. It mitigates risk through instructions that treat external content as untrusted and restrict the agent to read-only capabilities.
- Ingestion points: Code diffs and repository file content inspected during the review process (SKILL.md).
- Boundary markers: The skill contains explicit anti-pattern guidance against treating external logs or tool output as trusted instructions.
- Capability inventory: Shell execution is limited to read-only git operations; file system access is restricted to reading code; network operations and file modifications are explicitly prohibited.
- Sanitization: The skill relies on agent guidelines and the 'Anti-Patterns' section to manage untrusted data processing.
Audit Metadata