mcp-builder

Guides building MCP servers in Python or TypeScript with evaluation tooling.

1|Updated Mar 5, 2026
One-click install
npx skills add https://github.com/Clay-HHK/claude-skills --skill mcp-builder-clay-hhk
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: mcp-builder
Source: https://github.com/Clay-HHK/claude-skills/tree/main/mcp-builder
Command: npx skills add https://github.com/Clay-HHK/claude-skills --skill mcp-builder-clay-hhk

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 testing, which is difficult to get right without structured guidance. ## Core Features & Use Cases - Four-Phase Development Workflow: Covers research and planning, implementation, review and testing, and evaluation creation for MCP servers. - Language-Specific Guides: Provides reference documentation for both Python (FastMCP) and Node/TypeScript (MCP SDK) implementations with naming conventions, Zod/Pydantic schemas, and quality checklists. - Evaluation Harness: Includes scripts to run XML-based evaluation suites against MCP servers over stdio, SSE, or streamable HTTP transports and generate accuracy reports. - Use Case: When integrating an external API like GitHub or Slack with an LLM, use this Skill to design well-named tools, implement the server, and verify it with 10 complex read-only evaluation questions. ## Quick Start Help me build an MCP server in TypeScript that wraps the GitHub API with well-designed tools and evaluations.

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, build with npm run build, and test using the MCP Inspector via npx @modelcontextprotocol/inspector.

How do I build an MCP server in Python?▼

Use the Python SDK with FastMCP, defining tools with the @mcp.tool decorator and Pydantic models for input validation. Name the server {service}_mcp, verify syntax with python -m py_compile, and test with the MCP Inspector.

Should I use stdio or streamable HTTP for my MCP server?▼

Use stdio for local integrations and command-line tools running as subprocesses of a single client. Use streamable HTTP for remote servers, web services, and multi-client scenarios; SSE is deprecated in favor of streamable HTTP.

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 reports accuracy, tool call counts, and agent feedback on tool design.

What makes a good MCP tool name and description?▼

Use snake_case with a service prefix like github_create_issue to avoid conflicts with other servers. Descriptions should precisely state functionality, document parameters and return schemas, and include annotations like readOnlyHint and destructiveHint.

Why do MCP evaluation questions need stable answers?▼

Answers verified by direct string comparison must not change over time, so questions should target historical or closed data rather than current state like open issue counts. This ensures evaluation results remain reproducible across runs.