security-review
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest and analyze external code, which serves as a surface for indirect prompt injection attacks. Malicious instructions could be embedded in the code files being audited to deceive the agent or bypass the review process.
- Ingestion points: Processes user-provided source code, git diffs, and specific subsystems (identified in SKILL.md under 'Step 1: Scope the review').
- Boundary markers: The skill does not define specific text delimiters for the code input, but it does instruct the agent to ignore generic summaries and focus strictly on finding 'actual input source' and 'actual validation'.
- Capability inventory: The skill utilizes file read access and orchestrates sub-agents for exploration; it can suggest filing GitHub issues but requires explicit user confirmation before taking external actions.
- Sanitization: Procedural sanitization is enforced through the 'Success Criteria' and 'Core Principles', which mandate that every finding must cite a specific 'file:line' and provide a concrete exploit scenario, reducing the risk of the agent following misleading instructions found in comments or documentation.
Audit Metadata