The Agent Skills Directory

[SAFE]: The skill documents how to interact with a repository-specific security mechanism (protected-state-check) and defines a compliant workflow for resolving commit blocks without bypassing security controls.
[SAFE]: It explicitly warns against common bypass techniques, such as using unauthorized environment variables or git verification flags ('--no-verify'), and directs the agent to report bugs through proper infrastructure channels.
[SAFE]: The included diagnostic Node.js command uses local project libraries to verify file status and does not involve external network connections, data exfiltration, or execution of untrusted remote content.

protected-commit