README - mcp-handoff-server by dazeb

🤝 MCP Handoff Server

A Model Context Protocol (MCP) server that helps AI agents hand off work to each other with structured documentation and progress tracking.

✨ What it does

When AI agents need to pass work between each other, this server provides:

📝 Structured handoff documents with templates
🔄 Progress tracking from start to completion
📁 Automatic organization of active and archived handoffs
🔍 Easy searching and filtering of past work

🚀 Quick Start

Just run it with npx - no installation needed:

# Start in MCP mode (for MCP clients)
npx -y mcp-handoff-server

# Start HTTP server (for testing/direct API access)
npx -y mcp-handoff-server --mode http

That's it! The server automatically creates all needed folders and templates.

📋 Basic Usage

For MCP Clients

Add to your MCP client configuration:

{
  "mcpServers": {
    "handoff": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-handoff-server"
      ]
    }
  }
}

For HTTP Testing

# Start server
npx -y mcp-handoff-server --mode http

# Test it works
curl http://localhost:3001/health

🛠️ Available Tools

The server provides these MCP tools for AI agents:

graph LR
    A[📝 create_handoff] --> B[📖 read_handoff]
    B --> C[🔄 update_handoff]
    C --> D[✅ complete_handoff]
    D --> E[📦 archive_handoff]

    F[📋 list_handoffs] --> B

    style A fill:#e1f5fe
    style C fill:#f3e5f5
    style D fill:#e8f5e8
    style E fill:#fff3e0
    style F fill:#fce4ec

Tool Functions:

create_handoff - Start a new handoff document
read_handoff - Read an existing handoff
update_handoff - Add progress updates
complete_handoff - Mark work as finished
archive_handoff - Move completed work to archive
list_handoffs - Find and filter handoffs

📖 Example: Creating a Handoff

# Start the server
npx -y mcp-handoff-server --mode http

# Create a new handoff
curl -X POST http://localhost:3001/mcp \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "create_handoff",
    "params": {
      "type": "quick",
      "initialData": {
        "date": "2025-06-30",
        "time": "14:30 UTC",
        "currentState": {
          "workingOn": "Building user login",
          "status": "50% complete",
          "nextStep": "Add password validation"
        },
        "environmentStatus": {
          "details": {
            "Server": "✅",
            "Database": "✅"
          }
        }
      }
    }
  }'

🔧 Command Options

npx -y mcp-handoff-server [options]

Options:
  --mode <mode>        'mcp' or 'http' (default: mcp)
  --port <port>        HTTP port (default: 3001)
  --handoff-root <dir> Storage directory (default: ./handoff-system)
  --help              Show help
  --version           Show version

🔄 How It Works

Simple Workflow

Create a handoff when starting work
Update progress as you work
Complete when finished
Archive for future reference

graph TD
    A[🤖 Agent Starts Work] --> B{New Work?}
    B -->|Yes| C[📝 create_handoff]
    B -->|No| D[📖 read_handoff]

    C --> E[📁 Active Handoff]
    D --> E

    E --> F[🔄 update_handoff]
    F --> G{Work Done?}

    G -->|No| F
    G -->|Yes| H[✅ complete_handoff]

    H --> I[📦 archive_handoff]
    I --> J[🗄️ Archived]

    style C fill:#e1f5fe
    style F fill:#f3e5f5
    style H fill:#e8f5e8
    style I fill:#fff3e0

File Organization

The server automatically organizes everything in folders:

handoff-system/active/ - Current work
handoff-system/archive/ - Completed work
handoff-system/templates/ - Document templates

🎯 Two Types of Handoffs

📋 Standard Handoff - For complex work with detailed context ⚡ Quick Handoff - For simple updates and brief transitions

🏷️ Status Indicators

✅ Working - Everything good
⚠️ Warning - Some issues but not blocked
❌ Error - Problems that need fixing

🛠️ Development

Want to contribute or run locally?

# Clone and install
git clone <repository-url>
cd mcp-handoff-server
npm install

# Run in development
npm run dev

# Build for production
npm run build

📄 License

MIT License - feel free to use this in your projects!

🆘 Need Help?

Issues: GitHub Issues
MCP Protocol: Model Context Protocol Docs

Built for seamless AI agent collaboration 🤖✨