simplify
Pass
Audited by Gen Agent Trust Hub on Mar 30, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted data from the local repository via
git diff. This creates an indirect prompt injection surface where instructions embedded in code comments could attempt to influence the review agents. This is an inherent risk for any code-analysis tool and is addressed by the agent's internal safety guardrails. - Ingestion points:
SKILL.md(Phase 1, viagit diffor recently modified files) - Boundary markers: None explicitly defined for the diff content
- Capability inventory: The skill has file-writing capabilities to fix identified issues (Phase 3)
- Sanitization: None specified for the ingested diff data
- [COMMAND_EXECUTION]: The skill uses the
git diffcommand to gather context about recent code changes. This is a read-only operation and is standard for development-oriented skills.
Audit Metadata