commit
Pass
Audited by Gen Agent Trust Hub on Jul 2, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns or security risks were identified in the skill's instructions or workflow.
- [DATA_EXFILTRATION]: The skill explicitly instructs the agent to avoid staging or committing sensitive information such as secrets, credentials, and .env files, which is a defensive security best practice.
- [PROMPT_INJECTION]: The skill processes external data through git diffs and commit history, representing a surface for indirect prompt injection. However, the risk is mitigated by the specific scope of generating commit messages and the instruction to stop if the changes are not understood.
Audit Metadata