find-bugs
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [Indirect Prompt Injection] (SAFE): The skill processes untrusted code changes from a local repository as part of its auditing function. While this presents an indirect prompt injection surface, it is consistent with the intended behavior of a security analysis tool.
- Ingestion points: SKILL.md Phase 1 instructions use
git diffand direct file reads to ingest code into the context. - Boundary markers: Absent; the skill does not currently use specific delimiters for the audited content.
- Capability inventory: Command execution is limited to local read-only operations via
gitandgh. - Sanitization: Absent; the skill relies on the agent's reasoning to distinguish between code and potential malicious instructions.
Audit Metadata