The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is designed to execute a sequence of Git commands to manage code changes.
Evidence: The "Auto mode" triggers shell commands including git status, git add ., git commit -m "...", git log -1 --stat, and git push -u origin [branch-name].
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface where malicious data in the local environment could influence agent behavior during commit message generation.
Ingestion points: The skill reads the output of git status and processes user-provided strings for commit metadata (type, scope, subject).
Boundary markers: The skill does not define specific delimiters or instructions to ignore embedded commands within the files being staged or the generated commit messages.
Capability inventory: The skill has the capability to write to the local Git database (git commit) and transmit data over the network to a remote repository (git push).
Sanitization: No explicit sanitization or escaping is performed on the commit subject, although a character limit (50) is suggested.

git-commit