shap

Model interpretability and explainability using SHAP (SHapley Additive exPlanations). Feature importance, dependence plots, interaction effects, and fairness analysis for any black-box model.

Quality

58%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./scientific-skills/Data Analysis/shap/SKILL.md

Quality

Content

50%Weight 40%Scale 1-5

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

The skill body is rich with executable, actionable code but is held back by verbosity (theory re-explanation and redundant reference descriptions), workflows lacking validation checkpoints, and progressive-disclosure references that point to files which do not exist.

Suggestions

Trim the Key Concepts and Reference Documentation sections; move theory to references/theory.md and stop narrating reference-file contents inline.

Add explicit validation checkpoints (e.g., verify explainer output shape, sanity-check SHAP value sums against prediction-baseline) with feedback loops in the workflows.

Create the referenced bundle files (explainers.md, plots.md, workflows.md, theory.md) or remove the dangling references so progressive disclosure resolves to real files.

Dimension	Reasoning	Score
Conciseness	The body is noticeably verbose: the Key Concepts section explains SHAP-value theory Claude already knows, and the Reference Documentation section narrates the contents of each (nonexistent) reference file, adding padded sections beyond what earns its tokens.	2 / 5
Actionability	Provides substantial executable Python across quick start, common patterns, optimization, and integration examples that are mostly copy-paste ready, with minor gaps such as placeholder variable names and a dubious shap.plots.bar({...}) dict call.	4 / 5
Workflow Clarity	Six workflows list clear numbered sequences, but none include explicit validation checkpoints or validate->fix->retry feedback loops, so workflow clarity is capped at 3 per the batch/destructive-operations guideline.	3 / 5
Progressive Disclosure	References are clearly signaled ('See references/explainers.md', a Reference Documentation section, and Usage Guidelines), but the referenced files do not actually exist in the bundle and substantial content that belongs in separate files (theory, troubleshooting, integration) is inlined.	3 / 5
	Total	12 / 20 Passed

Description

66%Weight 40%Scale 1-5

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 and distinct, naming SHAP and several concrete capabilities with good natural trigger terms. Its main weakness is the absence of an explicit 'Use when...' clause, which caps completeness at 3.

Suggestions

Add an explicit 'Use when...' clause naming concrete user triggers (e.g., 'Use when explaining model predictions, generating SHAP plots, or analyzing feature importance/bias').

Reframe capabilities as verb-driven actions (e.g., 'Compute feature importance, generate dependence plots, analyze interaction effects') to lift specificity toward 5.

Include common user phrasings like 'explain my model' or 'why did my model make this prediction' to broaden trigger-term coverage.

Dimension	Reasoning	Score
Specificity	Names the SHAP domain and lists several concrete capabilities ('Feature importance, dependence plots, interaction effects, and fairness analysis'), but these are noun-phrase outputs rather than the verb-driven actions of the anchor-5 example, leaving minor coverage gaps.	4 / 5
Completeness	The description gives a clear 'what' (interpretability/explainability via SHAP) but contains no 'Use when...' clause or equivalent explicit trigger guidance, so completeness is capped at 3 per the rubric guideline.	3 / 5
Trigger Term Quality	Includes natural terms a user would say ('SHAP', 'feature importance', 'dependence plots', 'interaction effects', 'fairness analysis', 'black-box model'); a few common phrasings like 'explain my model' or 'why did my model predict' are absent.	4 / 5
Distinctiveness Conflict Risk	The SHAP-specific niche ('Model interpretability and explainability using SHAP') is clearly distinct with minimal conflict risk, though there is minor overlap with broader interpretability skills (LIME, general XAI).	4 / 5
	Total	15 / 20 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 (582 lines); consider splitting into references/ and linking	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning
referenced_paths_exist	Referenced path issues: 11 missing	Warning

	Total	13 / 16 Passed

Repository: aipoch/medical-research-skills
Path: scientific-skills/Data Analysis/shap/SKILL.md
Commit: f5ef65b

Reviewed: ~8 hours 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.