🎮 Game Dev Supercharger

Universal Claude Code MCP plugin for professional 2D/3D Unity game development

Transform your Unity game development workflow with AI-powered expert validation, automated asset generation, and performance profiling—all integrated directly into Claude Code.

✨ Features

🤖 Multi-Agent Expert Panel

Get expert validation before making critical decisions:

• GameDesignExpert: Player psychology, game feel, engagement loops
• UnityArchitect: Unity patterns, architecture, best practices
• PerformanceEngineer: Optimization, frame budgets, profiling
• VisualDirector: Art direction, visual cohesion, asset quality

🎨 Automated Asset Generation

Generate game-ready assets with AI:

• 2D Assets: Gemini (500/day free), Leonardo.ai, FLUX
• 3D Models: Meshy Pro, Hunyuan3D
• Batch Processing: n8n workflow templates included

⚡ Unity MCP Integration

Control Unity directly from Claude Code:

• Execute up to 25 commands in parallel
• Scene composition and asset management
• Real-time Unity Editor manipulation

📊 Performance Validation

Platform-specific performance profiling:

• Mobile: 60 FPS, 100 draw calls target
• Web: 30 FPS, 50 draw calls target
• Desktop: 60 FPS, 500 draw calls target
• Actionable optimization suggestions

🚀 Quick Start

Prerequisites

1. Unity MCP - Download from Asset Store
2. Python 3.10+ with uv package manager
3. Claude Code CLI installed

Installation

Method 1: Claude Code Plugin (Recommended)

# Install as a Claude Code plugin
/plugin marketplace add https://github.com/nategarelik/game-dev-supercharger
/plugin install game-dev-supercharger@game-dev-supercharger-dev

# Configure API keys in your environment
export ANTHROPIC_API_KEY="your-key-here"
export GEMINI_API_KEY="your-key-here"  # Optional
export LEONARDO_API_KEY="your-key-here"  # Optional
export MESHY_API_KEY="your-key-here"  # Optional

# Restart Claude Code to load the plugin

Method 2: Local Development

# Clone and test locally
git clone https://github.com/nategarelik/game-dev-supercharger.git
cd game-dev-supercharger

# Add local marketplace for testing
/plugin marketplace add /path/to/game-dev-supercharger
/plugin install game-dev-supercharger@game-dev-supercharger-dev

# Restart Claude Code

Method 3: Standalone MCP Server

# For advanced users who want to run the MCP server independently
git clone https://github.com/nategarelik/game-dev-supercharger.git
cd game-dev-supercharger
./install.sh

# Configure API keys
cp env.example .env
# Edit .env with your keys

# Add to Claude Code manually
claude mcp add --scope user GameDevSupercharger \
  -- uv --directory $(pwd)/mcp-server run server.py

API Keys Required

Add these to your environment variables (Windows: System Properties, macOS/Linux: ~/.bashrc or ~/.zshrc):

• ANTHROPIC_API_KEY (required): Expert panel functionality
• GEMINI_API_KEY (optional): Free 2D asset generation (500/day)
• LEONARDO_API_KEY (optional): Premium 2D assets (150 tokens/day free)
• MESHY_API_KEY (optional): 3D model generation ($20/mo)

📖 Usage Examples

Expert Panel Validation

spawn_expert_panel(
  decision="Use Physics2D for arcade baseball ball physics",
  experts=["GameDesignExpert", "UnityArchitect", "PerformanceEngineer"],
  context="Building retro arcade baseball targeting mobile and web"
)

Generate Game Assets

# Generate 2D sprites
generate_2d_assets(
  prompt="Baseball player sprite, batting pose, front view",
  style="pixel art, 16-bit, retro",
  count=4,
  tool="gemini"
)

# Generate 3D models
generate_3d_assets(
  prompt="Baseball stadium bleachers, weathered wood",
  type="text_to_3d",
  tool="meshy"
)

Unity Batch Commands

unity_batch_command(
  commands=[
    {"tool": "manage_scene", "action": "create", "name": "Level1"},
    {"tool": "manage_gameobject", "action": "create",
     "name": "Player", "components": ["Rigidbody2D", "BoxCollider2D"]},
    {"tool": "manage_material", "action": "create",
     "name": "PlayerMat", "color": "#FF0000"}
  ]
)

Performance Validation

validate_performance(
  platform="mobile",
  target_fps=60
)

🤖 Specialized AI Agents

The plugin includes 5 specialized agents, each with deep expertise and access to specific tools. Claude automatically invokes the right agent based on your request:

1. Expert Consultant 🎓

Validates decisions by consulting a panel of 4 AI experts (GameDesign, Unity, Performance, Visual).

Activates when you ask: "Should I use X or Y?", "Is this a good approach?"

Example:

You: "Should I use Physics2D or custom physics for my arcade game?"

Expert Consultant: "Let me consult the panel..."

• GameDesignExpert: Physics2D with arcade tweaks for predictable fun
• UnityArchitect: Battle-tested, integrates well with Unity systems
• PerformanceEngineer: Excellent mobile performance (< 3ms/frame)

Recommendation: Unity Physics2D + custom arcade force layer ✅

2. Asset Generator 🎨

Creates 2D sprites and 3D models using AI (Gemini, Leonardo, Meshy).

Activates when you ask: "Create sprites for...", "Generate a 3D model of..."

Example:

You: "I need player sprites for 4 teams"

Asset Generator: "I'll use Leonardo for style consistency across teams..." [Generates 16 sprite variations - 4 per team with team colors] "Here are your options. Should I have VisualDirector validate coherence?"

3. Performance Profiler ⚡

Profiles your game and identifies bottlenecks with actionable optimizations.

Activates when you ask: "Is my game performant?", "Check performance"

Example:

You: "Check if this runs well on mobile"

Performance Profiler: [Runs Unity Profiler] Result: ❌ FAIL - 247 draw calls (target: 100)

Fix: Create Sprite Atlas (saves ~150 calls), enable Static Batching (~30 calls) [Specific, actionable steps with expected impact]

4. Unity Builder 🔧

Rapidly builds Unity scenes with batch commands (up to 25 operations in parallel).

Activates when you ask: "Create a scene with...", "Set up GameObjects for..."

Example:

You: "Create a baseball field with all bases"

Unity Builder: [Executes 18 commands in one batch]

• Baseball diamond with 4 bases (collision detection)
• Pitcher's mound at regulation position
• Outfield wall, orthographic camera "Scene created in < 2 seconds! Ready for sprites?"

5. Game Architect 🏗️

Orchestrates all agents to take features from concept → implementation.

Activates when you ask: "Build a [feature]", "Plan implementation of..."

Example:

You: "Add a power-up system"

Game Architect: Phase 1: Expert validation → ScriptableObject pattern approved ✅ Phase 2: Generate 8 power-up icons with Asset Generator ✅ Phase 3: Build scene structure with Unity Builder ✅ Phase 4: Performance validation → 60 FPS maintained ✅

"Complete! Power-up system ready with expert-validated architecture."

💬 Natural Conversation Flow

Just talk naturally - Claude invokes the right agent automatically:

• "Should I use X?" → Expert Consultant
• "Create sprites" → Asset Generator
• "Is it fast?" → Performance Profiler
• "Build the scene" → Unity Builder
• "Add [feature]" → Game Architect (coordinates all agents)

📁 Project Structure

game-dev-supercharger/
├── .claude-plugin/      # Claude Code plugin manifests
│   ├── plugin.json      # Plugin metadata and MCP config
│   └── marketplace.json # Development marketplace
├── agents/              # Specialized AI agents (Claude Code)
│   ├── expert-consultant.md   # Decision validation agent
│   ├── asset-generator.md     # 2D/3D asset creation agent
│   ├── performance-profiler.md # Performance validation agent
│   ├── unity-builder.md       # Scene construction agent
│   ├── game-architect.md      # Orchestrator agent
│   ├── base_expert.py         # Expert panel base class
│   ├── game_design_expert.py  # Game design expert
│   ├── unity_architect.py     # Unity architecture expert
│   ├── performance_engineer.py # Performance expert
│   └── visual_director.py     # Visual direction expert
├── mcp-server/          # MCP server implementation
│   ├── server.py        # Main server with tool routing
│   └── config.py        # Configuration and environment
├── workflows/           # Asset generation pipeline
│   ├── asset_generation.py
│   └── n8n_templates/   # n8n workflow templates
├── tools/               # Performance validation
│   ├── performance_validator.py
│   └── performance_report.py
├── docs/                # Comprehensive documentation
│   ├── GUIDE.md         # Usage guide
│   ├── EXAMPLES.md      # Real-world examples
│   └── ARCHITECTURE.md  # System architecture
└── pyproject.toml       # Python dependencies

📚 Documentation

• Usage Guide: Complete feature documentation
• Examples: 5 real-world scenarios
• Architecture: System design and data flow

🛠️ Technology Stack

• Python 3.10+: Core implementation
• Anthropic SDK: Expert agent AI
• Unity MCP: Unity Editor integration
• httpx: Async HTTP client
• Gemini/Leonardo/Meshy: Asset generation APIs

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

📄 License

MIT License - see LICENSE file for details

🙏 Acknowledgments

Built with Claude Code and Claude Sonnet 4.5

• Inspired by professional game dev workflows
• Unity MCP integration enables direct editor control
• Multi-agent validation prevents costly architectural mistakes

📬 Support

• Issues: GitHub Issues
• Discussions: GitHub Discussions

---

Made with 🎮 by Retro Baseball Dev