model-evaluation-patterns
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill contains legitimate code patterns for machine learning evaluation and data analysis. It uses standard, well-maintained scientific libraries such as Scikit-Learn, Pandas, and NumPy for its primary functionality.
- [EXTERNAL_DOWNLOADS]: The skill fetches evaluation metrics like BERTScore, BLEU, and ROUGE from Hugging Face's official registry using the
evaluatelibrary. Hugging Face is a well-known service for machine learning assets and models.
Audit Metadata