code-review
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is a collection of instructional markdown files designed to guide an agent through a structured code review process. No malicious code, obfuscation, or unauthorized data access patterns were detected.
- [INDIRECT_PROMPT_INJECTION]: As a code review tool, the skill naturally ingests untrusted source code and VCS metadata, creating a surface for indirect prompt injection. However, the skill proactively mitigates this risk by including explicit instructions to delimit untrusted content and validate agent outputs.
- Ingestion points: Pull request descriptions and code diffs retrieved via git, gh, and glab as described in SKILL.md.
- Boundary markers: Present; guidelines require clear separation between trusted instructions and untrusted retrieved text.
- Capability inventory: The skill utilizes read-only VCS tools; there are no identified capabilities for arbitrary command execution or external network writes.
- Sanitization: The instructions mandate validation of model-generated output before any execution or storage.
Audit Metadata