The Agent Skills Directory

[SAFE]: The skill manages internal state and workflow logic for a development process. No security threats such as exfiltration, remote code execution, or harmful injections were identified.
[SAFE]: File access is restricted to the .gm/ directory, which is used for tracking project-specific progress and task rows.
[SAFE]: The skill implements a recursion bound via .gm/.gm-continue-count to prevent infinite loops during automated workflows.
[SAFE]: Project-specific commit attribution rules are defined for the 'anentrypoint' vendor resources, which is consistent with legitimate development practices.

gm-continue