The Agent Skills Directory

[SAFE]: The skill provides a set of instructions for an agent to evaluate code implementations based on a 5-criteria framework. It uses standard markdown formatting and does not contain any malicious patterns such as credential theft, network exfiltration, or unauthorized command execution.
[SAFE]: No external dependencies or remote code execution patterns were found. The skill operates entirely on the code implementations and data provided within the agent's existing context.
[SAFE]: The instructional language used (e.g., 'REQUIRED', 'CRITICAL') is appropriate for defining strict output constraints and formatting requirements; it does not attempt to bypass safety filters, jailbreak the model, or override system-level agent guidelines.

judge