What problem does it solve? Building Model Context Protocol servers requires navigating an evolving SDK API, choosing between stdio and HTTP transports, and correctly registering tools, resources, and prompts. This Skill provides the patterns and guidance to implement MCP servers without guessing at current API signatures. ## Core Features & Use Cases - Tool, Resource, and Prompt Registration: Register callable tools, read-only resources, and parameterized prompt templates using the @modelcontextprotocol/sdk with Zod schema validation. - Transport Selection: Configure stdio transport for local clients like Claude Desktop or Streamable HTTP for remote clients like Cursor and cloud deployments. - Use Case: You need to expose an internal API to AI assistants. Use this Skill to scaffold an MCP server, define Zod-validated input schemas for each tool, and connect it via stdio so Claude Desktop can invoke your endpoints. ## Quick Start Ask the AI to build an MCP server with a tool that queries your API, using Zod for input validation and stdio transport for Claude Desktop.