Design Patterns MCP Server

Instructions

Implementation Reference

• src/mcp-server.ts:350-379 (handler)

  Primary handler for the 'search_patterns' tool. Processes input arguments, performs semantic search using SemanticSearchService, formats results as markdown text content, and returns MCP-compatible response.
  private async handleSearchPatterns(args: any): Promise<any> { const query = { text: args.query, filters: args.filters || {}, options: { limit: args.limit || 10, includeMetadata: true, }, }; const results = await this.semanticSearch.search(query); return { content: [ { type: 'text', text: `Search results for "${args.query}":\n\n` + results .map( (result, index) => `${index + 1}. **${result.pattern.name}** (${result.pattern.category})\n` + ` Score: ${(result.score * 100).toFixed(1)}%\n` + ` Description: ${result.pattern.description}\n` ) .join('\n'), }, ], }; }
• src/mcp-server.ts:186-206 (registration)

  Tool registration in the list_tools response, defining the name, description, and input schema for 'search_patterns'.
  name: 'search_patterns', description: 'Search patterns by keyword or semantic similarity', inputSchema: { type: 'object', properties: { query: { type: 'string', description: 'Search query', }, searchType: { type: 'string', enum: ['keyword', 'semantic', 'hybrid'], default: 'hybrid', }, limit: { type: 'number', default: 10, }, }, required: ['query'], },
• src/lib/mcp-tools.ts:1016-1068 (handler)

  Alternative/complementary handler in MCPToolsHandler class supporting keyword, semantic, and hybrid search modes via adapters.
  private async handleSearchPatterns(args: any): Promise<any> { // Validate input if (!args.query || typeof args.query !== 'string' || args.query.length < 2) { throw new Error('Query must be at least 2 characters long'); } const searchType = args.search_type || 'hybrid'; const limit = Math.min(args.limit || 10, 50); try { let results: any[]; if (searchType === 'semantic') { // Use semantic search results = await this.config.semanticSearch.search(args.query, { limit, filters: args.category_filter ? { categories: [args.category_filter] } : undefined, }); } else if (searchType === 'keyword') { // Use keyword search results = await this.config.databaseManager.searchPatterns(args.query, { limit, filters: args.category_filter ? { categories: [args.category_filter] } : undefined, }); } else { // Hybrid search const semanticResults = await this.config.semanticSearch.search(args.query, { limit: Math.ceil(limit / 2), filters: args.category_filter ? { categories: [args.category_filter] } : undefined, }); const keywordResults = await this.config.databaseManager.searchPatterns(args.query, { limit: Math.ceil(limit / 2), filters: args.category_filter ? { categories: [args.category_filter] } : undefined, }); // Merge and deduplicate results results = this.mergeSearchResults(semanticResults, keywordResults, limit); } return { patterns: results, total_results: results.length, search_type: searchType, limit, timestamp: new Date().toISOString(), }; } catch (error) { throw new Error( `Pattern search failed: ${error instanceof Error ? error.message : String(error)}` ); } }
• src/lib/mcp-tools.ts:524-557 (registration)

  Tool registration in MCPToolsHandler.getTools(), including detailed input schema with validation constraints.
  name: 'search_patterns', description: 'Search patterns using keyword or semantic search', inputSchema: { type: 'object', properties: { query: { type: 'string', description: 'Search query', minLength: 2, maxLength: 500, }, search_type: { type: 'string', description: 'Type of search to perform', enum: ['keyword', 'semantic', 'hybrid'], default: 'hybrid', }, category_filter: { type: 'array', items: { type: 'string', }, description: 'Filter by pattern categories', }, limit: { type: 'number', description: 'Maximum number of results', minimum: 1, maximum: 50, default: 10, }, }, required: ['query'], },
• src/services/pattern-storage.ts:135-143 (helper)

  Database helper method for keyword-based pattern search using SQL LIKE queries.
  async searchPatterns(query: string): Promise<Pattern[]> { const searchTerm = `%${query}%`; const sql = ` SELECT * FROM patterns WHERE name LIKE ? OR description LIKE ? OR tags LIKE ? ORDER BY name `; return this.db.query<Pattern>(sql, [searchTerm, searchTerm, searchTerm]); }