Feature Orchestrator

Feature Orchestrator Skill

🚨 CRITICAL: This Skill Runs BEFORE Feature Implementation

Purpose: Enforce research-backed CI/CD workflow, prevent code duplication, ensure incremental delivery

When to invoke: Automatically when user requests implementing ANY new feature, component, or functionality

Auto-invoke triggers:

• "implement X feature"
• "add Y functionality"
• "build Z component"
• "create new X"
• "develop X feature"

What this skill does NOT apply to:

• Bug fixes ( "Incremental delivery enables organizations to have greater visibility, decreases risks faster, delivers value sooner."

Continuous Integration (Harness.io, 2024): > "Small problems are easier to fix than big problems, and frequent commits make bugs easier to identify."

Code Review (Microsoft Research): > "Reviewing 200–400 lines of code at a time detects up to 90% of defects."

DRY Principle (MetriDev, 2024): > "Abstraction and modularity prevent code duplication through identifying common patterns for reuse."

For complete research citations and detailed procedures: See REFERENCE.md

---

🎯 Four-Phase Workflow

Overview

| Phase | Purpose | Time | Output | |-------|---------|------|--------| | 1. Gap Analysis | Search for existing code | 2-3 min | Reuse opportunities identified | | 2. Planning | Break into increments | 3-5 min | Implementation plan with todos | | 3. Review Gate | Agent quality review | 1-2 min | Feedback incorporated | | 4. Execution | Build incrementally | Varies | Feature complete, tests passing |

---

Phase 1: Gap Analysis (MANDATORY)

Purpose: Search for existing implementations BEFORE writing any code

Research: DRY principle - 100 | Invoke critic-agent | | Security critical | Auth, payments | Always review | | User-facing UI | Any | Invoke ui-ux-designer | | Simple addition | 20 lines → data files

• Performance: React.memo() for animations, expensive components
• Imports: No file extensions, use path aliases (@/...)
• Accessibility: WCAG 2.1 AA compliance

File Size Monitoring:

# Check before editing
./scripts/check-file-size.sh [file-path]

# Thresholds:
# 150-200 lines → ⚠️  Plan extraction
# 200-300 lines → 🚨 Extract before adding
# 300+ lines    → 🛑 MUST refactor first

Detailed procedures: See REFERENCE.md → Phase 4

Verification (per increment):

• [ ] Implemented ONLY this increment's scope?
• [ ] All tests passing?
• [ ] Todo marked completed?

---

🔄 Complete Workflow Example

See EXAMPLES.md for:

• Complete auth system implementation (2.5 hours)
• Simple feature addition (5 minutes)
• Gap analysis preventing duplication
• Review gate catching security issues

---

🎯 Integration with CLAUDE.md Rules

This skill enforces ALL mandatory CLAUDE.md rules:

1. ✅ Search for existing code FIRST (Phase 1: Gap Analysis)
2. ✅ Plan before implementing (Phase 2: Implementation Planning)
3. ✅ Incremental implementation (Phase 4: Never break the build)
4. ✅ Test between increments (Phase 4: lint + type-check + test)
5. ✅ React.memo() for performance (Phase 4: Code quality rules)
6. ✅ Coordinate agents (Phase 3: Review gate)
7. ✅ TodoWrite tracking (Phase 2 + Phase 4)
8. ✅ Component size limits (Phase 4: File size monitoring)

---

🚨 Anti-Patterns This Skill Prevents

❌ What NOT to Do:

1. Starting without gap analysis → Skill forces search first
2. No implementation plan → Skill requires incremental plan
3. Large commits → Skill enforces small increments
4. Breaking the build → Skill tests after each increment
5. Skipping code review → Skill invokes agents automatically
6. Giant components (>300 lines) → Skill monitors file size
7. Missing optimization → Skill checks React.memo(), etc.

---

📋 Verification Checklist (Before Completing)

Before marking feature complete, verify:

• [ ] Phase 1: Gap analysis completed?
• [ ] Phase 1: Existing code reused where possible?
• [ ] Phase 2: Implementation plan created?
• [ ] Phase 2: TodoWrite tracking set up?
• [ ] Phase 3: Agent review completed (if needed)?
• [ ] Phase 4: All increments implemented?
• [ ] Phase 4: All todos marked complete?
• [ ] Phase 4: All tests passing?
• [ ] Phase 4: No breaking changes introduced?
• [ ] Phase 4: Performance optimized?
• [ ] Phase 4: Accessibility verified (WCAG 2.1 AA)?
• [ ] Phase 4: Component sizes within limits?

If any answer is NO, STOP and complete that phase.

---

🚀 Expected Outcomes

When this skill runs successfully:

1. ✅ No duplicate code - Gap analysis finds existing implementations
2. ✅ No redundancy - Reuses existing components/services
3. ✅ Well-planned - Clear roadmap with time estimates
4. ✅ Incrementally built - Small, testable steps
5. ✅ High quality - Agent reviews catch issues early
6. ✅ Well-tested - Tests run after each increment
7. ✅ Performant - Optimization applied automatically
8. ✅ Accessible - WCAG compliance checked
9. ✅ Maintainable - Component sizes controlled
10. ✅ Tracked - TodoWrite shows clear progress

---

📊 Success Metrics

Measure skill effectiveness:

• ✅ Zero duplicate implementations in last 10 features
• ✅ All features have implementation plans
• ✅ No surprise breaking changes in commits
• ✅ Test pass rate > 95% on first try
• ✅ Code review feedback declining over time
• ✅ Component sizes staying within limits
• ✅ Feature delivery time predictable (within 20% of estimate)

For detailed metrics and KPIs: See REFERENCE.md → Success Metrics

---

🔧 Automation Scripts

Validate increments:

# Linux/Mac
./scripts/validate-increment.sh

# Windows
./scripts/validate-increment.bat

Check file sizes:

./scripts/check-file-size.sh [file-path]

All scripts include:

• Research citations
• Clear pass/fail status
• Actionable recommendations
• Exit codes for CI/CD integration

---

📚 Supporting Documentation

For AI Assistants:

• REFERENCE.md - Complete research citations, detailed procedures, troubleshooting
• EXAMPLES.md - Real-world workflow examples, before/after scenarios
• resources/gap-analysis-template.md - Gap analysis structure
• resources/implementation-plan-template.md - Complete planning template
• resources/increment-checklist-template.md - Per-increment verification

Research Sources:

Academic:

• Google Research: "Modern Code Review" (9M reviews analyzed)
• Microsoft Research: "Expectations of Modern Code Review" (900+ devs surveyed)
• IEEE: "Continuous Integration Research" (meta-analysis)
• Scrum.org: "Incremental Delivery Research" (2024)

Industry:

• SonarQube: Code quality standards
• Atlassian: Trunk-based development
• Harness.io: CI/CD best practices
• MetriDev: Code duplication research

Books:

• "The DevOps Handbook" (Gene Kim)
• "Accelerate" (Nicole Forsgren)
• "Building Maintainable Software" (O'Reilly)

---

💡 Tips for Maximum Effectiveness

1. Trust the process - Let all 4 phases run
2. Don't skip gap analysis - Even if you "know" nothing exists
3. Break steps small - 30-45 min increments ideal
4. Test frequently - After EVERY increment
5. Use agent reviews - They catch issues you miss
6. Keep todos updated - Reflects real progress
7. Use automation - Scripts save time and reduce errors

---

🔄 Integration with Other Skills

Works with:

• code-refactoring - Triggered when files exceed size limits
• ui-ux-audit - Invoked during Phase 3 for UI features
• devops-deployment - After Phase 4 for production deployment
• qa-testing - Referenced in Phase 2 for test strategy

---

Final Note

This skill is not optional. When user requests implementing any feature, this skill MUST run to enforce research-backed CI/CD workflow rules.

The goal: Ship high-quality features faster by preventing common mistakes through automation.

Research-backed. Industry-proven. Battle-tested.

---

Version History

• v1.0 (Oct 2025): Initial version based on CLAUDE.md
• v2.0 (Nov 2025): Added research citations, progressive disclosure, automation scripts, templates

For complete version history and detailed research: See REFERENCE.md

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: WomenDefiningAI
• Source: WomenDefiningAI/claude-code-skills
• License: MIT

Install and usage instructions live in the source repository linked above.