The Agent Skills Directory

[SAFE]: The skill provides conceptual architectural guidance and best practices for security controls in agentic systems, emphasizing human-in-the-loop for high-risk actions.- [SAFE]: No executable scripts or code are included in the skill; code snippets provided are for illustrative documentation purposes only.- [SAFE]: The content promotes robust security principles including 'fail-closed' logic, comprehensive audit logging, and risk classification of tool calls.- [SAFE]: No network operations, file system modifications, or credential handling are performed by the skill.

permission-gates