ml-data-leakage-guard

Detects and prevents data leakage in machine learning and mathematical modeling. Use after ML tasks involving data cleaning, feature engineering, data augmentation, algorithm development, normalization, missing value imputation, dimensionality reduction, feature selection, or time series modeling. Checks if features/statistics would be available at prediction time.

Quality

88%

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

Quality

Content

82%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 body is highly actionable with executable code throughout and a sensible reference structure, but it is lengthy and substantially duplicates content (10 inline patterns + 5 examples) that overlaps with the reference files. Tightening the inline material and deferring detail to the references would improve both conciseness and progressive disclosure.

Suggestions

Collapse the 10 inline 'Critical Leakage Patterns' into a compact summary table and move the full WRONG/CORRECT code pairs into references/leakage-patterns.md to reduce duplication and token load.

The 5 'Examples' re-cover normalization, imputation, target encoding, and temporal leakage already shown as patterns — keep only 1-2 non-overlapping examples (e.g., post-event feature, cross-validation) and point the rest to the references.

Present the detection procedure as an explicit numbered workflow with validation checkpoints (e.g., '1. List preprocessing steps → 2. Apply checklist → 3. Run Prediction-Time test → 4. Re-verify after fixes') to lift workflow_clarity above 4.

Dimension	Reasoning	Score
Conciseness	Mostly efficient and executable with minimal concept over-explanation, but ~250 lines of inline patterns plus 5 worked examples repeat several of the same cases (normalization, imputation, target encoding, temporal) that the reference files catalog, creating noticeable redundancy.	4 / 5
Actionability	Fully executable copy-paste code for all 10 patterns and 5 worked examples, plus a concrete detection checklist and a 'Prediction Time' test with specific questions — covering common cases comprehensively.	5 / 5
Workflow Clarity	The pattern catalog, 8-item checklist, and prediction-time test provide a clear implicit sequence for auditing code, but the detection workflow is not presented as explicit numbered steps with validation checkpoints; checkpoints are present but implicit.	4 / 5
Progressive Disclosure	Good structure with a clearly signaled References section pointing to three real one-level-deep files (leakage-patterns.md, temporal-leakage.md, detection-strategies.md), but the bulk of the pattern catalog is inlined in SKILL.md and overlaps with references/leakage-patterns.md rather than being split out.	4 / 5
	Total	17 / 20 Passed

Description

95%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 strong, third-person, and well-structured with a clear 'what' and explicit 'Use after...' trigger clause covering many natural ML-preprocessing terms. Minor specificity gap prevents a perfect specificity score, but trigger coverage, completeness, and distinctiveness are all excellent.

Dimension	Reasoning	Score
Specificity	Names the domain ('data leakage in machine learning and mathematical modeling') and concrete actions ('Detects and prevents', 'Checks if features/statistics would be available at prediction time'), with minor gaps in coverage rather than a full enumeration of every action.	4 / 5
Completeness	Explicitly answers both 'what' ('Detects and prevents data leakage... Checks if features/statistics would be available at prediction time') and 'when' ('Use after ML tasks involving...') with concrete trigger phrases.	5 / 5
Trigger Term Quality	Comprehensive natural keyword coverage ('data cleaning', 'feature engineering', 'data augmentation', 'normalization', 'missing value imputation', 'dimensionality reduction', 'feature selection', 'time series modeling') — terms users would actually say when needing this skill.	5 / 5
Distinctiveness Conflict Risk	Clear niche (data leakage in ML/mathematical modeling) with distinct triggers and a precise 'Checks if features/statistics would be available at prediction time' framing that minimizes conflict with other ML skills.	5 / 5
	Total	19 / 20 Passed

Validation

100%

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: foryourhealth111-pixel/Vibe-Skills
Path: bundled/skills/ml-data-leakage-guard/SKILL.md
Commit: d5ae560

Reviewed: 1 day 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.