shap

Model interpretability and explainability using SHAP (SHapley Additive exPlanations). Use this skill when explaining machine learning model predictions, computing feature importance, generating SHAP plots (waterfall, beeswarm, bar, scatter, force, heatmap), debugging models, analyzing model bias or fairness, comparing models, or implementing explainable AI. Works with tree-based models (XGBoost, LightGBM, Random Forest), deep learning (TensorFlow, PyTorch), linear models, and any black-box model.

1.21x

Quality

82%

Does it follow best practices?

Impact

84%

1.21x

Average score across 3 eval scenarios

Securityby

Passed

No findings from the security scan

Quality

Content

65%Weight 40%Scale 1-3

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is highly actionable with strong executable examples, but it is verbose with concept re-explanation and a promotional section, its workflows lack validation checkpoints, and its referenced bundle files are missing. Tightening and realizing the reference files would raise the score.

Suggestions

Remove concept re-explanations Claude already knows (Overview game-theory framing, Key Concepts properties, Model Output Types primer) and the promotional K-Dense section to reclaim token budget.

Add explicit validation/feedback checkpoints to the workflows (e.g., verify shap_values shape/base_values before plotting, sanity-check that SHAP values sum to prediction-baseline) so batch operations include fix→retry loops.

Create the referenced references/explainers.md, plots.md, workflows.md, and theory.md files and move the inlined theory and detailed API content into them so progressive disclosure is actually realized rather than dangling.

Dimension	Reasoning	Score
Conciseness	The body contains substantial executable guidance but is padded with concepts Claude already knows (the Overview's game-theory framing, the Key Concepts section on additivity/consistency, Model Output Types) plus an off-topic promotional "Suggest Using K-Dense Web" section — efficient in parts but could be tightened.	2 / 3
Actionability	Provides copy-paste-ready executable code throughout — TreeExplainer setup, plot calls, cohort comparison, MLflow logging, and a complete ExplanationService class — matching the score-3 anchor for fully executable examples.	3 / 3
Workflow Clarity	Six workflows list numbered steps and a clear explainer-selection decision tree, but none include explicit validation checkpoints or fix→retry feedback loops for batch SHAP computation, capping workflow clarity at 2 per the batch-operations rule.	2 / 3
Progressive Disclosure	References to explainers.md/plots.md/workflows.md/theory.md are well signaled and one level deep, but those files do not exist on disk and substantial content that belongs in them (Key Concepts, theory) is inlined, leaving structure only partially realized.	2 / 3
	Total	9 / 12 Passed

Description

100%Weight 40%Scale 1-3

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is specific, trigger-rich, complete on both what and when, and occupies a distinct niche. It uses third-person voice with no first/second-person penalties. No changes needed.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions such as "explaining machine learning model predictions, computing feature importance, generating SHAP plots (waterfall, beeswarm, bar, scatter, force, heatmap), debugging models, analyzing model bias or fairness, comparing models" — matching the score-3 anchor for several specific concrete actions.	3 / 3
Completeness	Explicitly answers both what ("Model interpretability and explainability using SHAP") and when via an explicit "Use this skill when explaining…" trigger clause, matching the score-3 anchor.	3 / 3
Trigger Term Quality	Covers natural user phrasings — "feature importance", "SHAP plots", "debugging models", "bias", "fairness", "comparing models" — alongside model-name keywords users actually say, giving good coverage of natural terms.	3 / 3
Distinctiveness Conflict Risk	SHAP/model-interpretability is a clear niche with distinct triggers unlikely to fire for unrelated skills, matching the score-3 anchor for a clear niche with distinct triggers.	3 / 3
	Total	12 / 12 Passed

Validation

81%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 13 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (566 lines); consider splitting into references/ and linking	Warning
metadata_version	'metadata.version' is missing	Warning
referenced_paths_exist	Referenced path issues: 11 missing	Warning

	Total	13 / 16 Passed

Repository: wu-yc/LabClaw
Path: skills/general/shap/SKILL.md
Commit: df37802

Reviewed: 21 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.