Hugging Face Evaluation Manager

The Hugging Face Evaluation Manager skill provides comprehensive tools for managing evaluation results in Hugging Face model cards. It supports extracting evaluation tables from README content, importing benchmark scores from Artificial Analysis API, and running custom model evaluations with vLLM or lighteval. Works with model-index metadata format for leaderboard integration and supports both inference providers and local GPU evaluation.