A document-driven development protocol with three decision gates that prevent solution-jumping and catch wrong foundations before you build on them.
Built by following the Development Protocol itself — a recursive self-completeness test.
RAW INTENT
│
▼
┌─────────────────────────────────────────────────┐
│ EXTRACTION (Step 0) │
│ Extract X (real problem) from Y (stated │
│ solution). 7 proven techniques: Goal Climb, │
│ No-Computer Check, Why-Tree, Contextual Probe, │
│ Mom Question, Problem Statement Wall, │
│ Job/Pain/Gain Map. │
└─────────────────────────────────────────────────┘
│
┌─────────────────────────────────────────────────┐
│ FUNDAMENTALS │
│ Identify one-way door decisions. Validate each │
│ with minimum prototype. Detect LLM bias. │
│ Then proceed to decomposition only when │
│ foundations are proven safe. │
└─────────────────────────────────────────────────┘
│
▼
│ DECOMPOSITION │
│ Cynefin classify → MECE tree → confirm each │
│ level → KNOWN / RESEARCH / PROTOTYPE routing │
└─────────────────────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────┐
│ AMBITION (research-interleaved) │
│ Each round: answer → research → ask → ... │
│ Tightening loop until convergence │
└─────────────────────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────┐
│ LANDSCAPE (research + Cynefin) │
└─────────────────────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────┐
│ VALIDATION (prototype gate) │
│ KILL / PIVOT / COMMIT based on evidence │
└─────────────────────────────────────────────────┘
│ (if COMMIT)
▼
┌─────────────────────────────────────────────────┐
│ SPECIFICATION │
│ + Design for Change section (v3) │
└─────────────────────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────┐
│ EXECUTOR (with 10 v3 rules) │
│ Interface, Test, Boundary, Size, Cycle, │
│ Appetite, AI, Abstraction, Dependency, Backlog │
└─────────────────────────────────────────────────┘
│
▼
┌─────────────────────────────────────────────────┐
│ POLISH → EXPLAINER → SPEC_SYNC │
│ → REVIEW → ship │
└─────────────────────────────────────────────────┘
Prep-sequence: EXTRACTION -> FUNDAMENTALS -> DECOMPOSITION -> AMBITION -> LANDSCAPE -> VALIDATION -> SPECIFICATION -> EXECUTOR -> POLISH -> EXPLAINER -> SPEC_SYNC -> REVIEW -> ship
See REVIEW.md for the independence protocol and fixed checklist.
| EXTRACTION.md | Gate 0 - Extract X (real problem) from Y (stated solution), 7 proven techniques |
| FUNDAMENTALS.md | Identify one-way door decisions, validate with minimum prototype, detect LLM bias |
| DECOMPOSITION.md | Intent decomposition - Cynefin classify, MECE tree, KNOW/RESEARCH/PROTOTYPE routing |
| AMBITION.md | Gate 1 - Research-interleaved dialogue to clarify intent |
| LANDSCAPE.md | Research protocol - map what exists |
| VALIDATION.md | Gate 2 - rapid prototyping with KILL/PIVOT/COMMIT |
| SPECIFICATION.md | Locked plan-IS-spec template (14 sections) |
| EXECUTOR.md | AI execution handoff with autonomy levels |
| POLISH.md | Human final pass after AI execution |
| EXPLAINER.md | AI-generated code explainer for non-coder (docs/EXPLAINER.md) |
| SPEC_SYNC.md | Spec-to-code fidelity gate with research backing (docs/SPEC_SYNC.md) |
| RULES.md | Core protocol governance (v3.0.0 — 12 sections) |
| METHODOLOGY.md | Post-project retrospective |
| PLAYBOOK.md | Infrastructure checklist |
| REVIEW.md | Meta-review gate — independent agent audits protocol compliance |
| .omo/reviews/ | Review findings archive — one file per review run |
Generated by following the Development Protocol (v3.0.0) through its PREP PHASE on itself — producing a refined version that centers the prototyping gate as the key innovation. July 2026.