commit-it
Pass
Audited by Gen Agent Trust Hub on Jun 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes multiple shell commands to interact with the local git repository, including
git status,git diff,git commit,git reset, andgit apply. It also writes temporary patch files to/tmpto support hunk-level staging. - [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection because it ingests untrusted data from the local repository and processes it to generate commit messages.
- Ingestion points: File contents and differences are pulled into the agent's context via
git diffandgit diff --cachedinSKILL.md. - Boundary markers: The instructions lack delimiters or explicit warnings for the agent to ignore any natural language instructions that might be present within the diffs being analyzed.
- Capability inventory: The skill possesses the ability to write files to the system (
/tmp), apply patches (git apply), and commit changes to the repository history (git commit) as described inSKILL.md. - Sanitization: No sanitization, filtering, or validation is performed on the content of the diffs before the agent uses them to plan and execute commits.
Audit Metadata