mcp-builder

Guides building and evaluating MCP servers in Python or TypeScript.

Updated Feb 17, 2026
One-click install
npx skills add https://github.com/wolfmanstout/agent-skills --skill mcp-builder-wolfmanstout
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: mcp-builder
Source: https://github.com/wolfmanstout/agent-skills/tree/main/subtrees/skills/skills/mcp-builder
Command: npx skills add https://github.com/wolfmanstout/agent-skills --skill mcp-builder-wolfmanstout

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, review, 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 (Pydantic, @mcp.tool) and TypeScript (Zod, registerTool), plus MCP best practices for naming, pagination, transports, and security. - Evaluation Harness: Scripts to create 10-question XML evaluations and run them against your server via stdio, SSE, or HTTP transports using Claude. - Use Case: You need to expose your company's internal API to LLM agents. Use this Skill to design well-named tools with proper schemas, implement the server in TypeScript, then generate and run an evaluation to verify LLMs can answer realistic questions with it. ## Quick Start Use the mcp-builder skill to help me build an MCP server that wraps the GitHub API in 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 TypeScript?

Use the MCP TypeScript SDK's McpServer class and registerTool method with Zod schemas for input validation. Follow the naming convention {service}-mcp-server, provide title, description, inputSchema, and annotations for each tool, then build with tsc and test using MCP Inspector.

How do I build an MCP server in Python?

Use the Python SDK with FastMCP, registering tools via the @mcp.tool decorator and validating inputs with Pydantic models. Name your server {service}_mcp, verify syntax with python -m py_compile, and test with MCP Inspector.

Should I use stdio or streamable HTTP for my 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, and stdio servers must log to stderr rather than stdout.

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

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

What makes a good MCP tool description?

Tool descriptions should concisely state functionality, document parameters with constraints and examples, and specify return schemas. Use snake_case names with service prefixes like github_create_issue, and include annotations such as readOnlyHint and destructiveHint.

What are the limitations of MCP server evaluations?

Evaluation questions must be read-only, independent, and have stable answers verifiable by direct string comparison, so they cannot test write operations or dynamic state. Answers relying on current counts or mutable data are unsuitable.