Umbrella epic for CRUX Phase 3

The CRUX spec (docs/specifications/crux-competitive-research-ux-workflows.md) benchmarks aprender against every major ML/AI competitor by converting competitor verbs into 250 provable contracts (contracts/crux-*-v1.yaml).

Phase 1 (100% spec coverage) landed in PR #916 @ 960f7d5c3 — all 250 contracts authored + dogfooded through pv validate / pv lint / pmat comply check.

Phase 3 = implement the 81 stories currently marked missing. This epic tracks the 8 child M-group epics.

Children

• M1 REST / CLI parity (6 stories) — CRUX M1: REST / CLI parity — 6 missing stories #918
• M2 Distributed training (5 stories) — CRUX M2: Distributed training — 5 missing stories #919
• M3 OpenCLAW agent parity (20 stories) — CRUX M3: OpenCLAW agent parity — 20 missing stories #920
• M4 Observability (12 stories) — CRUX M4: Observability — 12 missing stories #921
• M5 Advanced quantization (10 stories) — CRUX M5: Advanced quantization — 10 missing stories #922
• M6 Alignment (6 stories) — CRUX M6: Alignment — 6 missing stories #923
• M7 Data depth (11 stories) — CRUX M7: Data depth — 11 missing stories #924
• M8 Deployment (11 stories) — CRUX M8: Deployment — 11 missing stories #925

Exit criteria

• All 81 contracts promoted from missing → supported
• apr qa --crux reports 250/250 PASS
• Master contract status block supported_count reaches 250

References

• Master contract: contracts/crux-competitive-research-ux-v1.yaml (phase_3_missing)
• Subspec: docs/specifications/crux-competitive-research-ux-workflows.md
• Evidence: evidence/crux/
• Phase 1 landing PR: docs(crux): CRUX competitive-research-UX spec — 250/250 contracts + OpenCLAW resolution #916
• Milestone: CRUX Phase 3