Mcp Builder

composiohq/awesome-claude-pluginsOfficial

This walks you through building MCP servers that let LLMs interact with external APIs and services, covering both Python (FastMCP) and TypeScript implementations. The guidance focuses on designing tools for actual agent workflows rather than just wrapping API endpoints. Think consolidating related operations into single high-impact tools, optimizing responses for limited context windows, and writing error messages that guide the LLM toward correct usage. It includes a four-phase approach from research (fetching live MCP protocol docs) through implementation, code review, and creating evaluation tests. The eval framework is smart: questions must be independent, read-only, and require multiple tool calls to answer, which forces you to validate that your server actually enables complex workflows.

GitHub stars: 1.7K
last updated: May 1, 2026

View on GitHub

Install

$ npx -y skills add composiohq/awesome-claude-plugins --skill mcp-builder --agent claude-code

Installs into .claude/skills of the current project.

Files

MCP Server Development Guide

Overview

To create high-quality MCP (Model Context Protocol) servers that enable LLMs to effectively interact with external services, use this skill. An MCP server provides tools that allow LLMs to access external services and APIs.

High-Level Workflow

Creating a high-quality MCP server involves four main phases:

Phase 1: Deep Research and Planning

Agent-Centric Design Principles

Build for Workflows, Not Just API Endpoints:

Don't simply wrap existing API endpoints - build thoughtful, high-impact workflow tools
Consolidate related operations (e.g., schedule_event that both checks availability and creates event)
Focus on tools that enable complete tasks, not just individual API calls

Optimize for Limited Context: