psychology

chevp-ai-framework

Lifecycle Gates Guidelines Commands Agents Templates
timeline The 3-step lifecycle

Context → Exploration → Production

Three sequential steps. Three quality gates. Each step produces defined artifacts — no step is skipped.

Per task, not per project. Every feature, bugfix, and refactor starts at Context.

1

Context

Understand the system and the problem before proposing solutions. Read, verify, ask — no code.

Output: Context-Plan, Problem Statement, Hypotheses, Risks, System Spec, Architecture, ADRs.

2

Exploration

Plan, prototype, validate. Two sub-modes — A (problem) and B (solution). No production code yet.

Output: Feature Plan/Spec, ADRs, UX Prototype, Challenger output, insights.md.

3

Production

Build, verify, ship. Execute the approved plan. No new plans, no scope changes.

Output: Production-Plan, code, validation result, commits, updated docs.

verified G1 between steps verified G2 between steps verified G3

Steps × Roles Matrix

Seven cross-cutting roles operate across all three steps

Role Context Exploration Production
SDLCProcess governance, scopeSpec/prototype governanceImplementation rules
AI-PlansSystem Spec, scope definitionFeature plan, acceptance criteriaPlan execution, finalization
UX-ToolingPrototype, preview loopVisual validation
DevOpsBuild, commit, CI/CD
Software-ArchitectureArchitecture, ADRsADRs for new decisionsPattern enforcement
Context-EngineeringCLAUDE.md, what to readStoring specs/plansUpdate CLAUDE.md, docs
ChallengerOptionalMandatory before G2Only on scope changes

AI Modes

The AI infers the mode — the human writes naturally

Mode 1

Context

Signals: "what does", "explain", "analyze", new task, ambiguous start.

Allowed: read/verify, ask questions, create CTX plan, produce System Spec.

Not allowed: change code, create feature plans, alter scope.

Mode 2

Exploration

Signals: "plan", "design", "prototype", "how should we".

Allowed: create plan/spec, write ADRs, iterate prototypes, document risks.

Not allowed: production code, expanding scope unilaterally.

Mode 3

Production

Signals: "implement", "build", "code", "execute the plan".

Allowed: execute approved plan, run tests, verify build, commit.

Not allowed: new plans, scope expansion, unplanned changes.

arrow_back Backward jumps allowed at any time. Forward jumps require a passed gate + human approval.

Production fallback: when implementation reveals the plan is incomplete, the AI must fall back to Exploration — not silently patch the plan.

The 6 Decisions per Step

Each step has exactly 6 discrete decisions the human owns — surfaced via AskUserQuestion

Toward G1

Context decisions

  1. Confirm Problem Statement
  2. Confirm Hypotheses
  3. Confirm Risks
  4. Approve System Spec & Architecture
  5. Confirm scope (in / out / uncertain)
  6. Pass G1 (approve / reject / revise)
Toward G2

Exploration decisions

  1. Choose exploration-mode A / B
  2. Approve Feature Plan/Spec
  3. Pick Challenger alternative
  4. Confirm prototype validation
  5. Confirm Kill Criteria + insights.md
  6. Pass G2 (approve / reject / revise)
Toward G3

Production decisions

  1. Approve start of PRD execution
  2. Approve scope change / fallback
  3. Approve commits / PR at boundaries
  4. Confirm acceptance criteria
  5. Confirm insights.md updated
  6. Pass G3 (approve / reject / revise)

When Steps May Be Abbreviated

No step is ever skipped — but each can be compressed

Scenario Allowed compression
Small bugfix (< 10 lines)Verbal CTX & EXP, one-liner PRD, no UX-Tooling. Context still read & verified.
Pure refactoringUX-Tooling omitted
Visual feature (UI, shader)Nothing — all plans must be written
Architecture decisionUX-Tooling omitted, ADR mandatory
Spike / explorationOnly Context + Exploration, no production code

When abbreviating, the AI must state which condition applies. Silent abbreviation is forbidden.

menu_book

Read the source

The canonical lifecycle definition, machine-readable.

LIFECYCLE.md open_in_new