The Agent Skills Directory

[SAFE]: The skill provides documentation, guidelines, and example implementations for LLM-as-a-Judge evaluation techniques without introducing security vulnerabilities.\n- [SAFE]: Python code provided in scripts/evaluation_example.py consists of benign mock examples for direct scoring, pairwise comparison, and rubric generation using only standard library utilities.\n- [SAFE]: All links and references point to legitimate external academic research or public repositories with no indicators of malicious intent or data exfiltration paths.

advanced-evaluation