---
description: Set up multi-agent collaboration framework (Claude, Gemini, Codex) with guardrails
argument-hint: [--quick] [--full]
---

# Multi-Agent Guide

Set up multi-agent collaboration framework for projects where multiple AI agents work together.

## Triggers
- "set up multi-agent", "agent guardrails", "multi-agent collaboration"

## Quick Setup (Recommended)

Rapid deployment with minimal questions:

1. Assess project structure
2. Ask which agents participate (Claude/Gemini/Codex/Human)
3. Create framework files
4. Customize ownership matrix

## Files Created

```
your-project/
├── .agent-state/
│   ├── tasks.yaml          # Task registry
│   └── locks.yaml          # Lock registry
├── tools/
│   └── check-ownership.py  # Ownership verification
├── MULTI_AGENT_FRAMEWORK.md  # Consolidated rules
├── GEMINI.md               # Sub-agent directive (if selected)
└── CODEX.md                # Sub-agent directive (if selected)
```

## Agent Hierarchy

```
            ┌─────────────────┐
            │   Claude Code   │
            │  (Lead Agent)   │
            └────────┬────────┘
                     │
      ┌──────────────┼──────────────┐
      v              v              v
┌──────────┐  ┌──────────┐  ┌──────────┐
│  Gemini  │  │  Codex   │  │  Human   │
│(Research)│  │ (Speed)  │  │ (Review) │
└──────────┘  └──────────┘  └──────────┘
```

## Commit Message Format

```
[Agent] type(scope): description

Examples:
[Claude] feat(core): implement new feature
[Gemini] docs(api): update API documentation
[Codex] test(models): add unit tests
```

## Post-Setup

1. Set agent identity: `export AGENT_AUTHOR=claude`
2. Review ownership matrix in `MULTI_AGENT_FRAMEWORK.md`
3. Install pre-commit hooks: `pre-commit install` (optional)

## Source
Full details: `/Users/ourdigital/Projects/our-claude-skills/custom-skills/91-multi-agent-guide/README.md`
Related commands: `/Users/ourdigital/Projects/our-claude-skills/custom-skills/91-multi-agent-guide/commands/`