🩺 Brain Health Audit
End-to-end pipeline check · catches every spot where fake data could poison the model · --
🔌 Stage 1: Data source
🧠 Stage 2: Feature extraction
📊 Stage 3: Outcome labeling
🎓 Stage 4: Training pipeline
🎯 Stage 5: Predictions & UI
💡 How to read this
Each stage of the brain pipeline depends on the previous one. If stage 1 (data source) is broken, every downstream stage produces garbage no matter how good the code is.

A healthy brain has every check green AND has at least 20 real trades labeled as wins/misses. A degraded brain has the pipeline working but lacks data. A broken brain has a fundamental data integrity issue — fix that before doing anything else.