|
| 1 | +# Lighthouse MCP Server |
| 2 | + |
| 3 | +A Model Context Protocol (MCP) server implementation that bridges AI agents with Lighthouse decentralized storage. This server provides AI-accessible tools for file storage, dataset management, and IPFS operations through a standardized protocol. |
| 4 | + |
| 5 | +## 🎯 Features |
| 6 | + |
| 7 | +- **MCP Protocol Compliance**: Full implementation of MCP specification |
| 8 | +- **Mock Lighthouse Operations**: Realistic file upload, download, and pinning simulations |
| 9 | +- **Dataset Management**: Create and manage file collections with metadata |
| 10 | +- **Tool Registry**: Dynamic tool registration and discovery system |
| 11 | +- **Request Validation**: Comprehensive input validation and sanitization |
| 12 | +- **Structured Logging**: Detailed operation logging for debugging and monitoring |
| 13 | +- **Performance Optimized**: Meets strict performance requirements (<2s startup, <500ms operations) |
| 14 | + |
| 15 | +## 📦 Installation |
| 16 | + |
| 17 | +```bash |
| 18 | +# Install dependencies |
| 19 | +pnpm install |
| 20 | + |
| 21 | +# Build the server |
| 22 | +pnpm run build |
| 23 | + |
| 24 | +# Run tests |
| 25 | +pnpm test |
| 26 | + |
| 27 | +# Run tests with coverage |
| 28 | +pnpm run test:coverage |
| 29 | +``` |
| 30 | + |
| 31 | +## 🚀 Quick Start |
| 32 | + |
| 33 | +### Running the Server |
| 34 | + |
| 35 | +```bash |
| 36 | +# Start with default configuration |
| 37 | +node dist/index.js |
| 38 | + |
| 39 | +# Start with custom log level |
| 40 | +node dist/index.js --log-level debug |
| 41 | + |
| 42 | +# Start with custom storage limit |
| 43 | +node dist/index.js --max-storage 2147483648 |
| 44 | + |
| 45 | +# View help |
| 46 | +node dist/index.js --help |
| 47 | +``` |
| 48 | + |
| 49 | +### Programmatic Usage |
| 50 | + |
| 51 | +```typescript |
| 52 | +import { LighthouseMCPServer } from '@lighthouse-tooling/mcp-server'; |
| 53 | + |
| 54 | +// Create server instance |
| 55 | +const server = new LighthouseMCPServer({ |
| 56 | + name: 'lighthouse-storage', |
| 57 | + version: '0.1.0', |
| 58 | + logLevel: 'info', |
| 59 | + maxStorageSize: 1024 * 1024 * 1024, // 1GB |
| 60 | + enableMetrics: true, |
| 61 | +}); |
| 62 | + |
| 63 | +// Start the server |
| 64 | +await server.start(); |
| 65 | + |
| 66 | +// Get server statistics |
| 67 | +const stats = server.getStats(); |
| 68 | +console.log('Server stats:', stats); |
| 69 | + |
| 70 | +// Graceful shutdown |
| 71 | +await server.stop(); |
| 72 | +``` |
| 73 | + |
| 74 | +## 🛠️ Available MCP Tools |
| 75 | + |
| 76 | +### 1. lighthouse_upload_file |
| 77 | + |
| 78 | +Upload a file to IPFS via Lighthouse with optional encryption. |
| 79 | + |
| 80 | +**Parameters:** |
| 81 | +- `filePath` (required): Path to the file to upload |
| 82 | +- `encrypt` (optional): Whether to encrypt the file |
| 83 | +- `accessConditions` (optional): Access control conditions |
| 84 | +- `tags` (optional): Tags for organization |
| 85 | + |
| 86 | +**Example:** |
| 87 | +```json |
| 88 | +{ |
| 89 | + "filePath": "/path/to/file.txt", |
| 90 | + "encrypt": true, |
| 91 | + "tags": ["dataset", "ml-model"] |
| 92 | +} |
| 93 | +``` |
| 94 | + |
| 95 | +### 2. lighthouse_create_dataset |
| 96 | + |
| 97 | +Create a managed dataset collection with metadata. |
| 98 | + |
| 99 | +**Parameters:** |
| 100 | +- `name` (required): Dataset name |
| 101 | +- `description` (optional): Dataset description |
| 102 | +- `files` (required): Array of file paths to include |
| 103 | +- `metadata` (optional): Additional metadata |
| 104 | +- `encrypt` (optional): Whether to encrypt the dataset |
| 105 | + |
| 106 | +**Example:** |
| 107 | +```json |
| 108 | +{ |
| 109 | + "name": "Training Dataset", |
| 110 | + "description": "ML training data", |
| 111 | + "files": ["/data/train.csv", "/data/validate.csv"], |
| 112 | + "metadata": { |
| 113 | + "author": "Data Team", |
| 114 | + "version": "1.0.0" |
| 115 | + } |
| 116 | +} |
| 117 | +``` |
| 118 | + |
| 119 | +### 3. lighthouse_fetch_file |
| 120 | + |
| 121 | +Download and optionally decrypt a file from Lighthouse. |
| 122 | + |
| 123 | +**Parameters:** |
| 124 | +- `cid` (required): IPFS CID of the file |
| 125 | +- `outputPath` (optional): Local path to save the file |
| 126 | +- `decrypt` (optional): Whether to decrypt the file |
| 127 | + |
| 128 | +**Example:** |
| 129 | +```json |
| 130 | +{ |
| 131 | + "cid": "QmYwAPJzv5CZsnA...", |
| 132 | + "outputPath": "/local/path/file.txt", |
| 133 | + "decrypt": true |
| 134 | +} |
| 135 | +``` |
| 136 | + |
| 137 | +## 🏗️ Architecture |
| 138 | + |
| 139 | +``` |
| 140 | +LighthouseMCPServer |
| 141 | +├── ToolRegistry # Tool management and execution |
| 142 | +├── MockLighthouseService # File operations (upload, fetch, pin) |
| 143 | +├── MockDatasetService # Dataset management |
| 144 | +├── Handlers |
| 145 | +│ ├── ListToolsHandler # Handle tools/list |
| 146 | +│ ├── CallToolHandler # Handle tools/call |
| 147 | +│ ├── ListResourcesHandler # Handle resources/list |
| 148 | +│ └── InitializeHandler # Handle initialize |
| 149 | +└── Utilities |
| 150 | + ├── RequestValidator # Input validation |
| 151 | + ├── ResponseBuilder # Response formatting |
| 152 | + └── CIDGenerator # Mock CID generation |
| 153 | +``` |
| 154 | + |
| 155 | +## 📊 Performance Metrics |
| 156 | + |
| 157 | +The server meets the following performance requirements: |
| 158 | + |
| 159 | +- **Server Startup**: < 2 seconds |
| 160 | +- **Tool Registration**: < 100ms per tool |
| 161 | +- **Mock Operations**: < 500ms per operation |
| 162 | +- **Memory Usage**: < 50MB |
| 163 | + |
| 164 | +## 🧪 Testing |
| 165 | + |
| 166 | +```bash |
| 167 | +# Run all tests |
| 168 | +pnpm test |
| 169 | + |
| 170 | +# Run with coverage |
| 171 | +pnpm run test:coverage |
| 172 | + |
| 173 | +# Run in watch mode |
| 174 | +pnpm run test:watch |
| 175 | +``` |
| 176 | + |
| 177 | +### Test Coverage |
| 178 | + |
| 179 | +The test suite includes: |
| 180 | +- **Unit Tests**: Individual component testing |
| 181 | +- **Integration Tests**: End-to-end workflow testing |
| 182 | +- **Performance Tests**: Metric validation |
| 183 | + |
| 184 | +Target: **>90% code coverage** ✅ |
| 185 | + |
| 186 | +## 📝 Configuration |
| 187 | + |
| 188 | +### Server Config Options |
| 189 | + |
| 190 | +```typescript |
| 191 | +interface ServerConfig { |
| 192 | + name: string; // Server name |
| 193 | + version: string; // Server version |
| 194 | + logLevel: 'debug' | 'info' | 'warn' | 'error'; |
| 195 | + maxStorageSize: number; // Max storage in bytes |
| 196 | + enableMetrics: boolean; // Enable metrics collection |
| 197 | + metricsInterval: number; // Metrics collection interval (ms) |
| 198 | +} |
| 199 | +``` |
| 200 | + |
| 201 | +### Environment Variables |
| 202 | + |
| 203 | +```bash |
| 204 | +# Log level |
| 205 | +LOG_LEVEL=info |
| 206 | + |
| 207 | +# Maximum storage size (bytes) |
| 208 | +MAX_STORAGE_SIZE=1073741824 |
| 209 | + |
| 210 | +# Enable metrics |
| 211 | +ENABLE_METRICS=true |
| 212 | +``` |
| 213 | + |
| 214 | +## 🔍 Logging |
| 215 | + |
| 216 | +The server uses structured logging with different log levels: |
| 217 | + |
| 218 | +```typescript |
| 219 | +// Log levels: debug, info, warn, error |
| 220 | +logger.info('Operation started', { operationId: '123' }); |
| 221 | +logger.error('Operation failed', error, { context: {...} }); |
| 222 | +``` |
| 223 | + |
| 224 | +## 🤝 Integration with AI Agents |
| 225 | + |
| 226 | +The server follows the MCP specification, making it compatible with: |
| 227 | + |
| 228 | +- **Cursor AI**: Direct integration via MCP |
| 229 | +- **Claude Desktop**: MCP server connection |
| 230 | +- **Custom AI Agents**: Any MCP-compliant client |
| 231 | + |
| 232 | +### Example Client Configuration |
| 233 | + |
| 234 | +```json |
| 235 | +{ |
| 236 | + "mcpServers": { |
| 237 | + "lighthouse-storage": { |
| 238 | + "command": "node", |
| 239 | + "args": ["/path/to/lighthouse-mcp-server/dist/index.js"], |
| 240 | + "env": { |
| 241 | + "LOG_LEVEL": "info" |
| 242 | + } |
| 243 | + } |
| 244 | + } |
| 245 | +} |
| 246 | +``` |
| 247 | + |
| 248 | +## 🛡️ Error Handling |
| 249 | + |
| 250 | +The server implements comprehensive error handling: |
| 251 | + |
| 252 | +- **Validation Errors**: Input parameter validation |
| 253 | +- **Not Found Errors**: Missing files or datasets |
| 254 | +- **Operation Errors**: Failed uploads or downloads |
| 255 | +- **System Errors**: Resource exhaustion |
| 256 | + |
| 257 | +All errors follow MCP error code standards. |
| 258 | + |
| 259 | +## 📚 API Documentation |
| 260 | + |
| 261 | +### Tool Registry API |
| 262 | + |
| 263 | +```typescript |
| 264 | +// Register a tool |
| 265 | +registry.register(toolDefinition, executor); |
| 266 | + |
| 267 | +// Execute a tool |
| 268 | +const result = await registry.executeTool(toolName, args); |
| 269 | + |
| 270 | +// List all tools |
| 271 | +const tools = registry.listTools(); |
| 272 | + |
| 273 | +// Get metrics |
| 274 | +const metrics = registry.getMetrics(); |
| 275 | +``` |
| 276 | + |
| 277 | +### Mock Service API |
| 278 | + |
| 279 | +```typescript |
| 280 | +// Upload file |
| 281 | +const result = await lighthouseService.uploadFile({ |
| 282 | + filePath: '/path/to/file', |
| 283 | + encrypt: true |
| 284 | +}); |
| 285 | + |
| 286 | +// Fetch file |
| 287 | +const file = await lighthouseService.fetchFile({ |
| 288 | + cid: 'QmYwAPJzv5CZsnA...' |
| 289 | +}); |
| 290 | + |
| 291 | +// Get storage stats |
| 292 | +const stats = lighthouseService.getStorageStats(); |
| 293 | +``` |
| 294 | + |
| 295 | +## 🔮 Future Enhancements |
| 296 | + |
| 297 | +- [ ] Integration with real Lighthouse SDK (Issue #11) |
| 298 | +- [ ] WebSocket support for real-time updates |
| 299 | +- [ ] Caching layer for improved performance |
| 300 | +- [ ] Batch operation support |
| 301 | +- [ ] Advanced access control management |
| 302 | + |
| 303 | +## 📄 License |
| 304 | + |
| 305 | +MIT License - see LICENSE file for details |
| 306 | + |
| 307 | +## 🤝 Contributing |
| 308 | + |
| 309 | +Contributions are welcome! Please read CONTRIBUTING.md for guidelines. |
| 310 | + |
| 311 | +## 📞 Support |
| 312 | + |
| 313 | +For issues and questions: |
| 314 | +- GitHub Issues: [lighthouse-agent-tooling/issues](https://github.com/Patrick-Ehimen/lighthouse-agent-tooling/issues) |
| 315 | +- Documentation: See `/apps/docs/TECHNICAL_PROPOSAL.md` |
| 316 | + |
| 317 | +--- |
| 318 | + |
| 319 | +**Built with ❤️ for the Lighthouse ecosystem** |
| 320 | + |
0 commit comments