The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes standard Git commands (git diff --cached and git commit) to perform its stated function of generating and applying commit messages.
[DATA_EXPOSURE_AND_EXFILTRATION]: The skill reads staged file changes to generate context for the commit message. This operation is localized to the user's repository and is necessary for the skill's primary purpose.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data (staged code changes) which could potentially contain malicious instructions. However, the skill implements a human-in-the-loop confirmation step before any permanent action (committing), which effectively mitigates the risk of the agent executing unintended instructions derived from the code diff.

commit