Guide MCP server development with TypeScript or Python SDKs.
npx skills add https://github.com/yashs33244/my-mac-claude --skill mcp-builder-yashs33244
Please help me install this Agent Skill. Skill: mcp-builder Source: https://github.com/yashs33244/my-mac-claude/tree/main/skills/anthropic/mcp-builder Command: npx skills add https://github.com/yashs33244/my-mac-claude --skill mcp-builder-yashs33244
Building an MCP server that other agents can reliably use is hard because tool discovery, naming, schemas, context handling, pagination, and error messages must all work together for successful end-to-end tasks.
Use the mcp-builder skill to create an MCP server that exposes a third-party API as well-named tools with strong schemas and pagination, then verify it by writing an evaluation XML with 10 complex read-only QA pairs.
High-intent search queries and answers about installing and using this skill.
Building an MCP server with reliable pagination involves designing discoverable, workflow-oriented tools and implementing type-safe schemas. This ensures agents can process large datasets efficiently and complete end-to-end tasks without context overload.
The best way to define input schemas for Model Context Protocol tools is using type-safe validation libraries like Zod in TypeScript or Pydantic in Python. This ensures structured data returns and helps agents understand tool parameters accurately.
You create evaluations for MCP tools by generating stable, read-only evaluation questions with verifiable tasks. Writing an evaluation XML with complex read-only QA pairs tests whether an LLM can effectively use your tools for robust performance.
Yes, you can use FastMCP with Python to integrate external services via Model Context Protocol. It supports exposing third-party APIs as well-named tools with strong schemas, reliable pagination, and actionable error handling.
Choosing between stdio and streamable HTTP transport for an MCP server depends on your deployment context and integration requirements. The skill provides transport guidance to harden reliability and ensure consistent agent communication.
LLM agents fail to complete tasks when MCP tool discovery, naming, schemas, context handling, pagination, and error messages are not properly aligned. Balancing broad API coverage with workflow-oriented tools and actionable errors resolves this.
felix-howardCreate MCP servers for LLM integration with FastMCP and Node/TypeScript SDKs.
bobmatnycStandardize Model Context Protocol server evaluation with read-only, idempotent task testing.
dbx0Generate MCP servers for Python and TypeScript with validated tools.
wangx7Build MCP servers integrating REST APIs with Python FastMCP or TypeScript SDKs.
