mcp-builder

Guides building and evaluating MCP servers in Python or TypeScript.

3|1|Updated Nov 30, 2025
One-click install
npx skills add https://github.com/PALabs-v1/AI_friend --skill mcp-builder-palabs-v1
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: mcp-builder
Source: https://github.com/PALabs-v1/AI_friend/tree/main/.claude/skills/mcp-builder
Command: npx skills add https://github.com/PALabs-v1/AI_friend --skill mcp-builder-palabs-v1

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires anthropic, mcp, and includes scripts (resource) and references (resource) components.

What problem does it solve? Building MCP (Model Context Protocol) servers that LLMs can actually use effectively requires careful tool design, schema validation, transport selection, and evaluation—this Skill provides a structured four-phase workflow covering research, implementation, testing, and evaluation. ## Core Features & Use Cases - Four-Phase Development Workflow: Deep research and planning, implementation with FastMCP (Python) or MCP SDK (TypeScript), code review and testing, and evaluation creation. - Language-Specific Reference Guides: Detailed implementation patterns for Python and Node/TypeScript including Zod/Pydantic schemas, tool annotations, pagination, and error handling. - Evaluation Harness: Scripts to create XML-based QA evaluations and run them against MCP servers via stdio, SSE, or streamable HTTP transports. - Use Case: You need to expose the GitHub API to an LLM agent. Follow the workflow to design well-named tools, implement them with proper schemas and annotations, then generate 10 complex evaluation questions and measure how accurately an LLM answers them using your server. ## Quick Start Ask the AI to help you build an MCP server for your chosen API, specifying whether you want Python or TypeScript.

Frequently Asked Questions about mcp-builder

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I build an MCP server in Python or TypeScript?

Follow the four-phase workflow: research the MCP specification and target API, implement tools using FastMCP (Python) or the MCP TypeScript SDK with Zod schemas, test with the MCP Inspector, then create evaluations. Reference guides cover project structure, tool registration, and quality checklists for each language.

What transport should I use for an MCP server?

Use streamable HTTP for remote servers serving multiple clients, and stdio for local integrations and command-line tools. SSE is deprecated in favor of streamable HTTP. Stateless JSON over streamable HTTP is recommended for simpler scaling.

How do I test whether my MCP server works well with LLMs?

Create an XML evaluation file with 10 complex, read-only questions having single verifiable answers, then run scripts/evaluation.py against your server. The harness uses Claude to attempt each question and reports accuracy, tool call counts, and agent feedback on tool design.

What naming conventions should MCP tools follow?

Use snake_case with a service prefix in the format service_action_resource, such as github_create_issue or slack_send_message. Python servers follow the {service}_mcp pattern while TypeScript servers use {service}-mcp-server.

Why do MCP evaluation questions need stable answers?

Answers must not change over time because verification uses direct string comparison. Questions based on current state like open issue counts become invalid as data changes, so evaluations should target historical, closed data.