commit
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements standard version control utility functions. It uses local git commands to manage code changes, following best practices such as avoiding bulk staging (e.g., preventing the use of
git add -A). - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the local repository environment to generate commit metadata.
- Ingestion points: Data enters the agent context through the output of
git status,git diff, andgit branchcommands as specified inSKILL.md. - Boundary markers: The instructions do not employ specific delimiters to isolate external repository data from internal logic.
- Capability inventory: The skill uses the
Bashtool to executegit addandgit commitcommands based on processed data. - Sanitization: The skill requires a mandatory human-in-the-loop confirmation step before any permanent changes are made to the repository, allowing users to review the final commit message and staged file list.
Audit Metadata