The Agent Skills Directory

[SAFE]: The skill consists solely of behavioral instructions for the AI and contains no code, scripts, or external dependencies.
[SAFE]: The 'Auto-Clarity Exception' is a positive security design that ensures critical information like security warnings and destructive action confirmations are delivered clearly rather than in 'caveman' format.
[SAFE]: The frontmatter 'disable-model-invocation: true' is a platform configuration that restricts the agent's capabilities, further reducing the potential attack surface.

zj-caveman