catchup
Pass
Audited by Gen Agent Trust Hub on Jun 23, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill requires the ability to execute git commands to read uncommitted changes from the local repository.\n- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it processes untrusted content from file diffs which could contain malicious instructions.\n
- Ingestion points: Output from local git diff commands.\n
- Boundary markers: None specified to differentiate between file content and instructions.\n
- Capability inventory: Reading local file modifications via git diff.\n
- Sanitization: No sanitization or filtering of the diff content is performed.
Audit Metadata