The Agent Skills Directory

[SAFE]: The skill serves a legitimate development purpose by enforcing a standard for commit messages and branch naming. It utilizes standard command-line tools (git, gh) for their intended purposes without any signs of misuse.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data in the form of git diffs and status outputs to generate commit messages.
Ingestion points: git status and git diff output.
Boundary markers: None explicitly defined for isolating the content of the diff from the instructions.
Capability inventory: git add, git commit, and gh CLI commands.
Sanitization: The skill relies on the agent's internal logic to parse the diff and generate a summary, though it includes a verification step to ensure the message matches the staged changes.

conventional-commits