NeuronC2

Advanced Command & Control Framework powered by Model Context Protocol

Features · Architecture · Installation · Usage · API · Security

NeuronC2 is a Command & Control framework that integrates with the Model Context Protocol (MCP), enabling AI-powered agent management through natural language. Control agents, execute commands, capture screenshots, and manage deployments — all through your MCP-compatible AI client.

graph TD
    A[MCP Client] -->|MCP Protocol| B[NeuronC2 Server]
    B -->|HTTP/WebSocket| C[Agent]
    B -->|SQLite| D[Database]
    
    subgraph Server
        B1[MCP Handler] --> B2[HTTP Server]
        B2 --> B3[WebSocket Manager]
        B3 --> B4[Database Manager]
    end
    
    subgraph Agent
        C1[Activation] --> C2[Command Executor]
        C2 --> C3[Screenshot Capture]
    end

neuronc2/
├── client/               # Agent implementation
│   ├── client.go         # Windows agent with WebSocket reconnect
│   ├── deploy.bat        # Windows deployment script
│   └── deploy.sh         # Linux deployment script
├── cmd/neuronc2/         # Server entry point
│   └── main.go           # HTTP + MCP server bootstrap
├── config/
│   └── config.go         # Environment-based configuration
├── internal/
│   ├── agent/            # Agent session management
│   ├── auth/             # Token authentication
│   ├── database/         # SQLite data access layer
│   ├── mcptools/         # MCP tool definitions
│   ├── server/           # HTTP/WebSocket server
│   └── utils/            # Shared utilities
├── pkg/models/           # Exportable data models
└── assets/               # Logo and images

• Go 1.24+
• SQLite3
• Windows (agent target)

git clone https://github.com/3xploit666/neuronc2.1.git
cd neuronc2.1
go mod download
go build -o neuronc2 ./cmd/neuronc2/main.go
./neuronc2

The server starts on:

• HTTP :8080 — agent communication
• MCP stdio — AI client integration

Add to your MCP client config (e.g., claude_desktop_config.json):

{
  "mcpServers": {
    "neuronc2": {
      "command": "/path/to/neuronc2"
    }
  }
}

1. Generate a deployment token via MCP
2. Compile the agent:

# Windows
cd client && deploy.bat -Token "DEPLOY-xxxxxxxxxxxxx"

# Generate a deployment token
mcp.generate_deployment_token(
    notes="Production deployment",
    duration="7d",
    max_uses=5
)

# List connected agents
agents = mcp.list_agents()

# Execute command on agent
result = mcp.send_command(
    agent_id="agent-abc123",
    command="whoami"
)

1. Admin generates deployment token (with expiry + max uses)
2. Agent activates with token → receives unique API key
3. All communication authenticated via X-API-Key header
4. Tokens auto-expire and track usage count

• Use HTTPS in production (reverse proxy with TLS)
• Rotate API keys regularly
• Set short token expiration times
• Monitor command history for anomalies
• Deploy behind network segmentation

Activate a new agent with a deployment token.

// Request
{
  "token": "DEPLOY-xxxxxxxxxxxxx",
  "metadata": {
    "hostname": "DESKTOP-ABC",
    "username": "john",
    "os": "windows",
    "arch": "amd64"
  }
}

// Response
{
  "agent_id": "agent-abc123",
  "api_key": "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",
  "status": "activated"
}

WebSocket endpoint for agent communication. Requires X-API-Key header.

• Encrypted communications (TLS/mTLS)
• Web monitoring dashboard
• Advanced command history search
• Automatic agent update system
• Linux/macOS agent support

This tool is for educational purposes and authorized testing only. Users are responsible for complying with applicable laws. The author assumes no responsibility for misuse or damage caused by this software.

@3xploit666 · Blog