feat(02-02): make-train orchestrator + Makefile/gitignore (Pattern 6, D-REPRO-04)

feat(02-03): implement plots — confusion matrix, dual reliability grid, lead-time CDFs

feat(02-03): implement per-class ECE/Brier + eval_metrics builder

feat(02-04): PyOD IForest trainer + 95p threshold + per-class lead-time scorer

feat(02-01): inference adapter — predict_verdict + per-mode mask

feat(02-01): feature matrix assembly + calibrated LightGBM classifier

feat(02-04): normal-baseline synthetic generator + make synth-normal

feat(02-01): wire Phase 2 ML deps + master seed plumbing

feat(01-03): generate.py entry point + columnar Parquet writer

feat(01-03): 10 per-class state machines + GENERATORS registry

feat(01-03): wave 0 — model repo skeleton + 3 generator tests + Makefile