What transport mechanisms are covered in the guide?

The guide covers both Streamable HTTP for remote, scalable servers and stdio for local servers used in desktop environments or CLI-based workflows.

What languages does the MCP Builder skill support?

It provides detailed implementation patterns and guidance for both TypeScript (using the official MCP SDK) and Python (using FastMCP and the Python SDK).

Is this skill suitable for beginners building their first MCP server?

Absolutely. It provides a step-by-step workflow from research to deployment, including links to necessary documentation and project structure templates.

How does this skill improve MCP server quality?

It enforces industry best practices in tool naming, Zod/Pydantic schema design, and descriptive error handling, ensuring that LLMs can easily discover and reliably use your tools.

Does this include testing and evaluation tools?

Yes, it includes a dedicated phase for creating XML-based evaluation pairs to verify that the LLM can solve realistic, complex tasks using your server tools.

MCP Server Builder

Name: MCP Server Builder
Author: kyh0703

bykyh0703

0•

API開発

Guides the creation of high-quality Model Context Protocol (MCP) servers to bridge external services and APIs with AI models.

The mcp-builder skill provides a comprehensive framework for developing robust MCP servers in TypeScript and Python, ensuring LLMs can interact seamlessly with external data and tools. It leads developers through a structured four-phase workflow—Deep Research, Implementation, Review/Testing, and Evaluation—prioritizing discoverable tool naming, efficient context management, and actionable error handling. Whether building local stdio servers for personal use or remote HTTP implementations for production, this skill standardizes the integration of external APIs into the AI-native development lifecycle.

主な機能

010 GitHub stars

02Multi-language support for TypeScript (SDK) and Python (FastMCP)

03Detailed implementation patterns for core infrastructure and shared utilities

04Built-in evaluation framework for testing LLM tool-calling effectiveness

05Best practices for tool naming, schema definition, and pagination

06Standardized workflow from initial API research to final evaluation

ユースケース

01Creating local utilities that extend Claude's capabilities via stdio transport

02Building custom connectors for proprietary internal company APIs

03Developing public-facing MCP servers for SaaS platforms and developer tools

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add kyh0703/claude-plugins mcp-builder

For use in Claude.ai and ChatGPT

Download Skill