Rh Methodology Help

# Development Methodology — Quick Reference

Show the complete methodology workflow and available commands.

## Display this guide:

Development Methodology — Quick Reference ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

WORKFLOW PHASES:

Phase 0: Discovery /rh-project-kickoff Phase 1: Architecture /rh-generate-architecture /rh-generate-architecture-delta Phase 2: Sprint Planning /rh-generate-devspec /rh-generate-devspec-delta Phase 3: Implementation /rh-execute-sprint {N} 🔄 ralph-loop /rh-execute-all-sprints {range} 🔄 ralph-loop Phase 4: Quality /rh-quality-gate 🔄 ralph-loop Phase 5: Release /rh-release {version}

Status: /rh-sprint-status

Reverse Engineering: /rh-reverse-engineer {path} 🔄 ralph-loop

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

MODEL SELECTION:

Phase 0-2 (Thinking): Use Opus → /model opus Phase 3-5 (Doing): Use Sonnet → /model sonnet

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

TYPICAL NEW PROJECT FLOW:

1. /model opus
2. /rh-project-kickoff → Fill requirements from JIRA
3. /rh-generate-architecture → Generate architecture doc
4. /rh-generate-devspec → Generate sprint plan with TDD
5. /model sonnet
6. /rh-execute-all-sprints → Implement all sprints
7. /rh-quality-gate → Verify quality
8. /rh-release 1.0.0 → Build, tag, deploy

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

REVERSE ENGINEERING FLOW (existing codebase → documentation):

1. /model opus
2. /rh-reverse-engineer {path} → Analyzes code, asks questions, generates ALL 5 documents: CLAUDE.md, BUSINESS-LOGIC, ARCHITECTURE, DEVELOPMENT-SPEC, PROJECT-KICKOFF

Use cases: - Understand existing project in ultra-detail - Prepare for technology migration - Onboard new team members - Create specs for a rewrite

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

VERSION INCREMENT FLOW:

1. /model opus
2. /rh-generate-architecture-delta → Delta architecture
3. /rh-generate-devspec-delta → Delta sprint plan
4. /model sonnet
5. /rh-execute-all-sprints {range} → Implement new sprints
6. /rh-quality-gate → Verify quality
7. /rh-release {version} → Build, tag, deploy

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

TEMPLATES (in .claude/templates/):

METHODOLOGY.md Master workflow definition project-init/PROJECT-KICKOFF.md Requirements template architecture/ARCHITECTURE-.md Architecture templates development-spec/DEVELOPMENT-.md Sprint planning templates agents/AGENTS.md Agent roles and models

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

QUALITY TARGETS:

Core logic coverage: > 90% Adapter coverage: > 80% Global coverage: > 80% Lint issues: 0 actionable Security vulnerabilities: 0 Test race conditions: 0

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

RALPH WIGGUM LOOP (🔄):

Development commands auto-activate /ralph-wiggum:ralph-loop This ensures tasks complete even across context resets. Each command has a completion promise that MUST be true to exit.

/rh-execute-sprint → promise: SPRINT_DONE /rh-execute-all-sprints → promise: ALL_SPRINTS_DONE /rh-quality-gate → promise: QUALITY_PASSED /rh-reverse-engineer → promise: REVERSE_ENGINEERING_DONE

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

KEY PRINCIPLES:

✦ Architecture-first: No code until architecture is approved ✦ TDD mandatory: RED → GREEN → REFACTOR, always ✦ Regression protection: All previous tests must pass ✦ Traceability: Requirements → Components → Tests ✦ Context7: Verify all API patterns before use ✦ Opus for thinking, Sonnet for doing ✦ Ralph loop for dev tasks: iterate until genuinely done