What problem does it solve? Building MCP (Model Context Protocol) servers that LLMs can actually use effectively requires careful tool design, naming, pagination, error handling, and evaluation—this Skill provides a structured four-phase workflow to do it right. ## Core Features & Use Cases - Guided Development Workflow: Walks through research, implementation, review, and evaluation phases for MCP servers in Python (FastMCP) or TypeScript (MCP SDK). - Reference Documentation: Includes best practices, language-specific implementation guides, and an evaluation guide covering tool naming, transport selection, and security. - Evaluation Harness: Ships Python scripts to run XML-based QA evaluations against your MCP server via stdio, SSE, or HTTP transports using Claude. - Use Case: You want to expose the GitHub API to an LLM agent. Use this Skill to design well-named tools with Zod/Pydantic schemas, then generate 10 complex read-only evaluation questions and measure how accurately an agent answers them. ## Quick Start Use the mcp-builder skill to help me design and implement an MCP server for the Slack API in TypeScript, then create an evaluation file to test it.
