commit
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses shell commands to interact with Git (
git add,git commit). It includes a security best practice by explicitly instructing the agent to shell-quote each literal path and use--literal-pathspecs, which mitigates the risk of command injection through malicious filenames.- [INDIRECT_PROMPT_INJECTION]: The skill processes data in the form of git diffs and file contents which could theoretically contain instructions. - Ingestion points: Reads branch status, staged/unstaged diffs, and untracked contents in
SKILL.md. - Boundary markers: The agent is instructed to resolve scope from invocation arguments and authorized callers, and to ask the user before splitting concerns.
- Capability inventory: The agent can execute
gitcommands and read/write via tool calls. - Sanitization: Employs shell-quoting for path inputs and verification of commit SHAs after execution to ensure the result matches the intended scope.
Audit Metadata