mcp-builder

Create MCP servers with Python FastMCP or Node/TypeScript MCP SDK.

24|8|Updated Jan 14, 2026
One-click install
npx skills add https://github.com/vadimcomanescu/codex-skills --skill mcp-builder-vadimcomanescu
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: mcp-builder
Source: https://github.com/vadimcomanescu/codex-skills/tree/main/skills/.experimental/platform/mcp-builder
Command: npx skills add https://github.com/vadimcomanescu/codex-skills --skill mcp-builder-vadimcomanescu

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

This guide helps teams design and implement high-quality MCP (Model Context Protocol) servers so LLMs can reliably interact with external APIs and services through well-defined tools and transports.

Core Features & Use Cases

  • Design guidance: Balances API coverage and workflow tools, naming conventions, and actionable error messages to improve agent discoverability and resilience.
  • Language-specific guides: Concrete implementation patterns for Python (FastMCP) and Node/TypeScript (MCP SDK) including Pydantic/Zod schemas, transport selection, and structured outputs.
  • Testing & evaluation: Includes an evaluation harness and an evaluation guide to create read-only, verifiable tests that validate tool usefulness for LLMs.
  • Utilities & examples: Provides connection helpers, example scripts, and best practices for pagination, authentication, and secure error handling.

Quick Start

Implement your MCP server following the Phase 1–3 guides for your chosen language, then run the included evaluation harness with the provided evaluation XML to validate tools via stdio or HTTP.

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 for LLMs to interact with external APIs?

To build an MCP server for LLMs, you define well-defined tools and transports that enable the model to interact with external services. This Skill provides design guidance and implementation patterns for Python FastMCP and Node/TypeScript MCP SDK.

What is the best way to structure input and output schemas for an MCP server?

The best way to structure MCP server schemas is using Pydantic for Python FastMCP or Zod for Node/TypeScript MCP SDK. This ensures clear tool naming, structured responses, and actionable error messages for improved agent discoverability.

Does this guide support both stdio and streamable HTTP transports for MCP servers?

Yes, this guide supports both stdio and streamable HTTP transports for MCP servers. It provides best practices for transport selection, pagination, authentication, and secure error handling across the supported Python and Node/TypeScript implementations.

How do I verify that my MCP server tools are useful for LLMs?

You verify MCP server tools by running an included evaluation harness with provided evaluation XML. This creates read-only, verifiable tests that validate tool usefulness for LLMs via stdio or HTTP connections.

When do I need to use FastMCP versus the Node TypeScript MCP SDK?

You choose FastMCP for Python implementations or the Node TypeScript MCP SDK for JavaScript environments when building MCP servers. Both paths offer concrete patterns for API integration, workflow tools, and structured outputs.

Can I use this to add workflow tools and API integration to an existing LLM agent?

Yes, you can use this to add workflow tools and API integration to an existing LLM agent. It balances API coverage and workflow tool design to improve agent resilience and external service interaction.