The Agent Skills Directory

[SAFE]: The skill consists entirely of instructional markdown files (SKILL.md, examples.md, patterns.md, reference.md). It functions as a specialized knowledge base for the agent rather than a set of executable scripts.
[SAFE]: No executable code, shell scripts, or binary files are present in the skill directory. While the documentation contains Python code blocks and JSON examples, these are for illustrative purposes and templates to be followed by the user or the agent when designing evaluations, not for autonomous execution.
[SAFE]: The skill does not perform network operations, access sensitive file paths, or require elevated privileges. Its primary function is to guide the user in creating robust evaluation plans with a focus on observable behavior and safety checks.

skill-evaluation