The Agent Skills Directory

[SAFE]: The skill implements standard Git workflow automation for staging and committing code changes without suspicious system or network behavior.\n- [COMMAND_EXECUTION]: The skill utilizes local Git commands (git status, git add, git diff, git reset, git commit) to perform its primary function. Operations are strictly confined to the local repository environment.\n- [DATA_EXFILTRATION]: The guidelines explicitly prohibit staging sensitive information like environment variables, credentials, or private keys, which mitigates the risk of accidental data exposure.\n- [PROMPT_INJECTION]: The skill has an indirect prompt injection surface. 1. Ingestion points: Untrusted data enters the context via git diff --cached execution in SKILL.md. 2. Boundary markers: None are defined to separate code content from instructions. 3. Capability inventory: The agent can execute git add, git reset, and git commit commands. 4. Sanitization: There is no evidence of escaping or filtering diff content before it is analyzed to generate commit messages.

git-commit