Split train and tool simulator modules; mastery curriculum and grader workflow nudge.

Harden training parser and validation workflow

Rename files to DriftShield: notebooks, server module, kernel folder

Cleanup: drop legacy support-ops cruft now that DriftShield is the only collection

DriftShield Phase B: 4 new grader components, 5 new penalties, negation-aware forbidden phrases

feat: GRPO training stack (Qwen3 + Unsloth opt-in, curriculum, eval, audit)

fix: map grader task scores to open interval (0,1) for Phase 2

add expert task: GDPR compliance trap with misleading signals

fix: align inference.py env vars with spec and clean up repo

fix: match mandatory stdout format from hackathon spec

clean rewrite: shorter vars, no comments, same logic