shap

SUMMARY

What it does

This skill provides procedural guidance for explaining and auditing machine-learning predictions using the SHAP library (version 0.52.0). It covers selecting appropriate explainers and maskers, computing feature attributions via the modern shap.Explanation API, validating explanations through additivity checks, handling multi-output models, and producing local or global visualizations. The skill includes operating rules, installation instructions, a standard workflow, common task patterns, troubleshooting order, and references to detailed guides. It emphasizes explicit output selection, background data choice, and reporting limitations. The source is a Markdown file (SKILL.md) from the K-Dense-AI/scientific-agent-skills repository, inspected successfully.