MASTER PLAN — Complete Research Stack Remaining Work
Document ID: MASTER_PLAN_2026-04-19
Authority: AGENTS.md §9 — Research prioritization
Scope: All remaining P0, P1, P2, and Layer M work
Executive Summary
Current Status: 6 assignments complete, computational verification framework established
Remaining Work: Documentation updates, symbolic proof generalization, Layer M expansion
Estimated Effort: 4-6 weeks for complete closure
Phase 0: Documentation Reconciliation (1-2 days)
Problem
Documentation (RESEARCH_ROADMAP, MATH_MODEL_MAP) lists items as "OPEN" that are actually computationally verified with complete frameworks.
Action Items
| File |
Task |
Lines |
Status |
RESEARCH_ROADMAP.md |
Update P0 missingLinkODE status |
~338-339 |
⏳ Pending |
RESEARCH_ROADMAP.md |
Update P1 General Proofs section |
~339 |
⏳ Pending |
RESEARCH_ROADMAP.md |
Mark Key Lemmas as complete |
~370-376 |
⏳ Pending |
MATH_MODEL_MAP.md |
Update theorems 122-124 status |
~356-358 |
⏳ Pending |
MATH_MODEL_MAP.md |
Update theorem 131 status |
~365 |
⏳ Pending |
MATH_MODEL_MAP_BY_DOMAIN.md |
Sync with main map |
All |
⏳ Pending |
Deliverable
All documentation accurately reflects computationally verified + framework established status for completed work.
Phase 1: Symbolic Proof Generalization (3-4 weeks)
Current State
All P0/P1/P2 theorems have computational verification via native_decide.
Gap: Symbolic proofs for unbounded/general cases (no concrete values).
Symbolic Proof Targets
Tier 1: Critical for CAI Analysis (1 week)
| Theorem |
Domain |
Approach |
Difficulty |
caiBounds (162) upper bound |
AM-GM |
Real→Float transfer |
Medium |
rscuSumSynonymous symbolic |
Enumeration |
General fold proof |
Hard |
speciesBetterThanGeneric |
Entropy |
Real analysis |
Hard |
Tier 2: P1 Geometry (1-2 weeks)
| Theorem |
Domain |
Approach |
Difficulty |
tipCoordinateMassResonance general |
Diophantine |
Hyperbola parametrization |
Research |
tipCoordinateMirrorResonance general |
Integer arith |
Symmetric solution space |
Research |
fortyFiveLineFactorRevelation general |
Number theory |
Fermat's theorem formalization |
Research |
Tier 3: P0 ODE (1-2 weeks)
| Theorem |
Domain |
Approach |
Difficulty |
missingLinkODE general existence |
Analysis |
Picard iteration convergence |
Research |
missingLinkODEUniqueness symbolic |
Lipschitz |
Contraction mapping |
Hard |
Research Strategy
- Real→Float Transfer: Prove in ℝ, transfer to Float for representable values
- Q16_16 Alternative: Convert Float-heavy proofs to fixed-point (AGENTS.md compliance)
- Constructive Witnesses: Provide explicit algorithms (Euler, Fermat) rather than pure existence
Phase 2: Layer M Model Expansion (2-3 weeks)
Current State
- Layer M proven: 30 models (44% auto-proven)
- Remaining: 38 models from 68 total
- Approach:
native_decide for concrete values, pattern matching for families
Remaining Model Categories
Category A: Enumeration-Based (Week 1)
| Model Range |
Count |
Approach |
Status |
| Codon/AA mappings |
~8 |
Exhaustive enumeration |
Ready |
| Species variants |
~5 |
Computational per-species |
Ready |
| Start/stop variants |
~3 |
Case analysis |
Ready |
Category B: Float Inequalities (Week 2)
| Model Range |
Count |
Approach |
Status |
| CAI bounds variants |
~12 |
Real analysis proof |
Needs work |
| Entropy comparisons |
~6 |
AM-GM + monotonicity |
Needs work |
| Thermodynamic bounds |
~4 |
Conservation laws |
Ready |
Category C: Research-Level (Week 3)
| Model Range |
Count |
Approach |
Status |
| ODE existence variants |
~3 |
Euler/Picard framework |
Framework ready |
| Braid topology |
~2 |
Combinatorial |
Open |
| Factorization geometry |
~2 |
Number theory |
Open |
Phase 3: Infrastructure & Polish (1 week)
Code Quality
| Task |
File |
Effort |
Status |
Remove all sorry placeholders |
AVMR.lean |
1 day |
⏳ Critical |
Add #eval witnesses for all theorems |
AVMR.lean |
2 days |
⏳ Needed |
| Standardize theorem documentation |
AVMR.lean |
1 day |
⏳ Formatting |
| Lake build zero warnings |
All |
Ongoing |
🔄 Active |
Testing
| Task |
Coverage |
Status |
lake build passes |
100% |
✅ Current |
native_decide theorems verified |
All concrete |
✅ Current |
| Symbolic proof regression tests |
New proofs |
⏳ Needed |
Resource Allocation
Subagent Assignments (Proposed)
| Subagent |
Phase |
Focus Area |
Duration |
| Cascade |
Phase 0 |
Documentation reconciliation |
2 days |
| Ω (ODE) |
Phase 1, Tier 3 |
ODE general existence |
1-2 weeks |
| α₁ |
Phase 1, Tier 2 |
Hyperbola geometry |
1-2 weeks |
| α₂ |
Phase 1, Tier 2 |
Mirror resonance general |
1 week |
| α₃ |
Phase 1, Tier 2 |
Fermat factorization |
1-2 weeks |
| β₁ |
Phase 2, Cat A |
Enumeration proofs |
1 week |
| β₂ |
Phase 2, Cat B |
Float inequalities |
1-2 weeks |
| β₃ |
Phase 2, Cat C |
Research-level models |
1-2 weeks |
| δ (Docs) |
Phase 3 |
Code quality, #eval |
1 week |
Dependency Graph
Phase 0 (Docs)
│
├──────┬────────┬────────┐
│ │ │ │
Phase 1 Tier 1 Tier 2 Tier 3 (Symbolic - parallel)
│ │ │ │
└──────┴────────┴────────┘
│
Phase 2 (Layer M - parallel)
│
Phase 3 (Polish)
Risk Assessment
| Risk |
Probability |
Impact |
Mitigation |
| Symbolic proofs intractable |
Medium |
High |
Stay with computational + framework |
| Float→ℝ transfer fails |
Low |
Medium |
Use Q16_16 fixed-point |
| Lake build breaks |
Low |
High |
CI pipeline, incremental builds |
| Research-level proofs require external help |
High |
Medium |
Document gaps, flag for experts |
Success Criteria
Phase 0: All documentation reflects actual completion status
Phase 1: All P0/P1 theorems have either symbolic proof OR documented framework
Phase 2: Layer M reaches 70%+ auto-proven (47+ models from 68)
Phase 3: Zero sorry, all theorems have computational or symbolic proof
Timeline Summary
| Phase |
Duration |
Deliverable |
| 0: Documentation |
1-2 days |
Accurate status tracking |
| 1: Symbolic Generalization |
3-4 weeks |
Theorems with general proofs |
| 2: Layer M Expansion |
2-3 weeks |
47+ models proven |
| 3: Infrastructure |
1 week |
Code quality, zero sorry |
| Total |
6-8 weeks |
Research stack complete |
Immediate Next Steps
- Today: Begin Phase 0 — Update RESEARCH_ROADMAP.md to reflect actual status
- This Week: Assign subagents to Phase 1 tiers based on expertise
- Next Week: Begin Phase 2 Layer M expansion with enumeration-based models
Plan Version: 1.0
Created: 2026-04-19
Authority: AGENTS.md §9
Status: APPROVED FOR EXECUTION