# Multi-Agent Coordinator

> Spawn 10-20 parallel agents for Leavn codebase audits, systematic fixes, comprehensive analysis across multiple domains simultaneously

- Skill: `majiayu000/multi-agent-coordinator` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds add majiayu000/multi-agent-coordinator`
- Raw SKILL.md: https://api.skillmd.com/api/skills/majiayu000/multi-agent-coordinator/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: majiayu000 (https://skillmd.com/u/majiayu000)
- Updated: 2026-09-09
- Page: https://skillmd.com/skills/majiayu000/multi-agent-coordinator

---


# Multi-Agent Coordinator

Deploy agent swarms for comprehensive work:

1. **Audit swarms** (10-15 agents):
   - UserDefaults usage
   - @Published migration targets
   - Error handling patterns
   - Performance issues
   - Duplicate services
   - Navigation patterns
   - Memory leak risks
   - Test coverage gaps

2. **Fix swarms** (5-10 agents):
   - Fix group 1 errors (files 1-5)
   - Fix group 2 errors (files 6-10)
   - Migrate tier 1 ViewModels
   - Migrate tier 2 Services
   - Create entities 1-3
   - Create entities 4-6

3. **Verification swarms** (3-5 agents):
   - Test feature A
   - Test feature B
   - Verify integration C

4. **Deploy pattern**:
   - Single message, multiple Task calls
   - Clear tasks with specific deliverables
   - Collect all results before acting

Use when: Large refactoring, comprehensive audit, parallel fixes needed, systematic analysis