review
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill defines a structured process for auditing code changes against project-specific and industry-standard guidelines. All behavior is consistent with the stated purpose of code quality assurance.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data in the form of code changes and documentation, which presents a surface for indirect prompt injection.
- Ingestion points: Context is collected from changed files, pull request diffs, and project-specific guidance files such as CLAUDE.md or REVIEW.md.
- Boundary markers: The instructions do not specify explicit delimiters (like XML tags or markdown blocks) to isolate untrusted code from instructions, relying on the model's standard context window management.
- Capability inventory: The skill is designed for agents with file system read access to the local project workspace.
- Sanitization: No specific content filtering or sanitization of the reviewed code is mandated within the prompt.
Audit Metadata