find-bugs
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION] (SAFE): Employs git and gh CLI tools to gather repository state and code diffs. This is routine behavior for a development-oriented skill.
- [DATA_EXFILTRATION] (SAFE): No network exfiltration patterns detected. Interaction with GitHub via the gh CLI is limited to metadata retrieval and is considered a trusted source interaction for this use case.
- [PROMPT_INJECTION] (SAFE): No instructions found that attempt to bypass safety filters or override agent behavior.
- [INDIRECT_PROMPT_INJECTION] (SAFE): The skill analyzes untrusted code changes which could contain malicious instructions. (1) Ingestion points: git diff output and individual file reads in SKILL.md. (2) Boundary markers: Absent. (3) Capability inventory: Reading files, execution of git and gh; no write or network export capabilities. (4) Sanitization: None. (5) Severity: SAFE as this is the primary purpose of a code auditing tool.
Audit Metadata