GLM-4.6 MCP Server

Instructions

Implementation Reference

• src/glm-client.ts:68-122 (handler)

  Core implementation of the consult_architecture tool: constructs a specialized system prompt for architecture consultation and calls the GLM-4.6 API.
  async consultArchitecture(query: string, context?: string): Promise<string> { const systemPrompt = `You are an elite software architecture consultant specializing in enterprise-grade system design, scalability patterns, security architecture, and technical decision-making. Your expertise includes: - Distributed systems architecture and microservices design - Cloud-native patterns and containerization strategies - Database architecture and data modeling - API design (REST, GraphQL, gRPC) - Security architecture and threat modeling - Performance optimization and scalability - DevOps and CI/CD pipeline architecture - Modern frontend and backend frameworks - System integration patterns Provide concise, actionable architectural guidance with enterprise-grade best practices. Focus on technical accuracy, scalability, maintainability, and security.`; const messages: GLMMessage[] = [ { role: 'system', content: systemPrompt }, ]; if (context) { messages.push({ role: 'user', content: `Context:\n${context}\n\nArchitectural Query:\n${query}`, }); } else { messages.push({ role: 'user', content: query }); } const request: GLMRequest = { model: this.model, messages, temperature: 0.7, top_p: 0.9, max_tokens: 4096, stream: false, }; try { const response = await this.client.post<GLMResponse>('/chat/completions', request); if (!response.data.choices || response.data.choices.length === 0) { throw new Error('GLM-4.6 returned empty response'); } return response.data.choices[0].message.content; } catch (error) { if (axios.isAxiosError(error)) { const status = error.response?.status; const message = error.response?.data?.error?.message || error.message; throw new Error(`GLM-4.6 API Error (${status}): ${message}`); } throw error; } }
• src/index.ts:142-153 (handler)

  MCP server handler for callTool request: extracts parameters and delegates to GLMClient.consultArchitecture.
  case 'consult_architecture': { const { query, context } = args as { query: string; context?: string }; const response = await glmClient.consultArchitecture(query, context); return { content: [ { type: 'text', text: response, }, ], }; }
• src/index.ts:25-42 (registration)

  Registration of the consult_architecture tool in the MCP tools list, including description and input schema.
  { name: 'consult_architecture', description: 'Consult GLM-4.6 for expert software architecture guidance, system design patterns, scalability strategies, and technical decision-making. Use this for high-level architectural questions requiring deep technical expertise.', inputSchema: { type: 'object', properties: { query: { type: 'string', description: 'The architectural question or problem requiring expert consultation', }, context: { type: 'string', description: 'Optional additional context about the system, requirements, or constraints', }, }, required: ['query'], }, },
• src/index.ts:28-41 (schema)

  Input schema definition for the consult_architecture tool, specifying query (required) and optional context.
  inputSchema: { type: 'object', properties: { query: { type: 'string', description: 'The architectural question or problem requiring expert consultation', }, context: { type: 'string', description: 'Optional additional context about the system, requirements, or constraints', }, }, required: ['query'], },