SDOF Knowledge Base

# Alternative Implementation Without MCP SDK Dependency If you're experiencing issues with the `@modelcontextprotocol/sdk` dependency, this guide provides an alternative implementation approach that doesn't rely on the SDK. ## The Issue The error message indicates that the specified version of the MCP SDK (`@modelcontextprotocol/sdk@^0.1.0`) isn't available: ``` npm error code ETARGET npm error notarget No matching version found for @modelcontextprotocol/sdk@^0.1.0. ``` ## Solution Options ### Option 1: Use Wildcard Version (Already Applied) We've updated the package.json to use a wildcard version specifier: ```json "@modelcontextprotocol/sdk": "*" ``` This will accept any available version of the SDK. Try running `npm install` again with this change. ### Option 2: Implement MCP Server Without the SDK If the wildcard approach doesn't work, you can implement a simple MCP server without the SDK dependency. Here's how: 1. Create a new `server.js` file in the root of your project: ```javascript import express from 'express'; import { PlanAutoSaveService } from './src/services/plan-auto-save.service.js'; const app = express(); app.use(express.json()); const PORT = process.env.PORT || 3000; // Define the MCP tools const tools = [ { name: 'store_sdof_plan', description: 'Stores an SDOF plan in the knowledge base and filesystem', handler: async (inputs) => { try { const planService = PlanAutoSaveService.getInstance(); const { title, content, planType, tags = [] } = inputs; const result = await planService.savePlan(title, content, planType, tags); return { success: true, filePath: result.filePath, entryId: result.entryId, message: `Plan "${title}" successfully saved to ${result.filePath} and knowledge base` }; } catch (error) { return { success: false, error: error instanceof Error ? error.message : 'Unknown error occurred', message: 'Failed to save SDOF plan' }; } } }, { name: 'search_knowledge', description: 'Searches the knowledge base for relevant entries', handler: async (inputs) => { // Implement search functionality // This is a placeholder - you would need to implement the actual search logic return { success: true, results: [], message: 'Search completed' }; } } // Add other tools as needed ]; // MCP API endpoint app.post('/mcp/v1/tools/:toolName', async (req, res) => { const { toolName } = req.params; const inputs = req.body; const tool = tools.find(t => t.name === toolName); if (!tool) { return res.status(404).json({ success: false, error: `Tool '${toolName}' not found` }); } try { const result = await tool.handler(inputs); return res.json(result); } catch (error) { return res.status(500).json({ success: false, error: error instanceof Error ? error.message : 'Unknown error occurred' }); } }); // Start the server app.listen(PORT, () => { console.log(`SDOF Knowledge Base MCP server running on port ${PORT}`); }); ``` 2. Add a new start script to package.json: ```json "scripts": { "build": "tsc && node -e \"require('fs').chmodSync('build/index.js', '755')\"", "start": "node build/index.js", "start:alt": "node server.js", "dev": "ts-node src/index.ts", "test": "jest" } ``` 3. Update the MCP configuration for Roo to use this alternative server: ```json { "name": "sdof_knowledge_base", "type": "stdio", "command": "npm run start:alt", "cwd": "/path/to/your/project" } ``` ### Option 3: Use a Mock Implementation for Development If you're just testing the auto-save functionality, you can create a mock implementation: 1. Create a `mock-server.js` file: ```javascript import fs from 'fs'; import path from 'path'; // Mock implementation of plan saving const savePlan = (title, content, planType, tags = []) => { // Create sanitized filename from title const sanitizedTitle = title .toLowerCase() .replace(/[^a-z0-9]+/g, '-') .replace(/(^-|-$)/g, ''); // Get current date for folder organization const date = new Date(); const dateStr = date.toISOString().split('T')[0]; // Determine directory based on plan type let dirPath = path.join('docs', 'plans', planType + 's'); // Create directory if it doesn't exist if (!fs.existsSync(dirPath)) { fs.mkdirSync(dirPath, { recursive: true }); } // Create file path const fileName = `${dateStr}-${sanitizedTitle}.md`; const filePath = path.join(dirPath, fileName); // Write file fs.writeFileSync(filePath, content); return { filePath, entryId: `mock-${Date.now()}` }; }; // Parse command line arguments const args = process.argv.slice(2); if (args.length >= 3) { const [title, planType, contentFile] = args; const content = fs.readFileSync(contentFile, 'utf8'); const tags = args.slice(3); const result = savePlan(title, content, planType, tags); console.log(JSON.stringify(result, null, 2)); } else { console.error('Usage: node mock-server.js "Plan Title" planType contentFile.md [tag1 tag2 ...]'); } ``` 2. Use this from your custom mode by executing the command: ``` node mock-server.js "SDOF Exploration: Knowledge Base" exploration exploration-content.md knowledge-base vector-database ``` ## Choosing the Right Approach 1. **If you need full MCP compatibility**: Try Option 1 first, then Option 2 2. **If you just need the auto-save functionality**: Option 3 provides a simple solution 3. **If you want to contribute to the MCP ecosystem**: Consider checking the official MCP documentation for the correct SDK version ## Next Steps After implementing one of these alternatives: 1. Test the functionality by creating a sample SDOF plan 2. Verify that the plan is saved to the filesystem in the correct location 3. If using Options 1 or 2, test the MCP tool integration with your SDOF Orchestrator These alternatives should allow you to proceed with implementing the auto-save functionality even if there are issues with the MCP SDK dependency.