Agent skill · research science · aiskillstore
ml-antipattern-validator
Prevents 30+ critical AI/ML mistakes including data leakage, evaluation errors, training pitfalls, and deployment issues. Use when working with ML training, testing, model evaluation, or deployment.
Why this skill is useful
Adds executable validation scripts and specific patterns to prevent critical AI/ML mistakes that the AI wouldn't generate on its own.
What it needs
Requires python installed locally. About 2k tokens when loaded. Last updated 2026-08-07. 405 stars on the source repository.
What this skill does
ML Antipattern Validator Overview AI/ML 개발에서 30+ 안티패턴을 감지하고 방지하는 스킬입니다. Key Principle: Honest evaluation > Impressive metrics. When to Activate Automatic Triggers: ML training code (train.py, model training) Dataset preparation or splitting Model evaluation or testing Production deployment planning Manual Triggers: @validate-ml - Full validation @check-leakage - Data leakage detection @verify-eval - Evaluation methodology --- Pre-Implementation Checklist --- Critical Antipatterns Category 1: Data Leakage 🚨 1.1 Target Leakage 1.2 Temporal Leakage 1.3 Preprocessing Leakage 1.4 Group Leakage 1.5 Data Augmentation Leakage --- Category 2: Evaluation Mistakes ⚠️ 2.1 Testing on Training Data 2.2 Metric Misalignment 2.3 Accuracy Paradox 2.4 Invalid Time Series CV 2.5 Hyperparameter Tuning on Test Set --- Category 3: Training Pitfalls 🔧 3.1 Batch Norm Inference Error 3.2 Early Stopping Overfitting 3.3 Learning Rate Warmup 3.4 Class Imbalance --- Detection Patterns Leakage Detection Mode Check --- Validation Checklist Before deployment: [ ] No data leakage detected [ ] Test set never seen during training [ ] Metrics aligned with business objective [ ] model.eval() called for inference [ ] Class imbalance handled [ ] Covariate shift monitoring planned --- References 상세 예시 및 시나리오는 references/REFERENCE.md 참조.
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills aiskillstore/ml-antipattern-validator