The Agent Skills Directory

[SAFE]: The skill follows security best practices for AI agents by requiring explicit human approval via the AskUserQuestion tool before performing sensitive actions like git commit.
[SAFE]: No evidence of prompt injection, data exfiltration, or obfuscation was found. The skill's operations are localized to the user's filesystem and standard git commands.
[DATA_EXPOSURE]: The skill reads from CLAUDE.md and executes git status/diff/log to gather context for a commit. This is standard behavior for a version control tool and does not involve exfiltration to external domains.
[INDIRECT_PROMPT_INJECTION]: Like any skill that processes external content (in this case, git diffs and logs), there is a theoretical surface for indirect prompt injection if the repository content contains malicious instructions. However, the mandatory human review of the generated commit proposal and the manual confirmation gate (AskUserQuestion) are effective mitigations.

commit-repo