mcp-builder

Guides building, testing, and evaluating MCP servers in TypeScript or Python.

Updated Feb 23, 2026
One-click install
npx skills add https://github.com/mashharuki/vibekanban-gitworktree-sample --skill mcp-builder-mashharuki
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: mcp-builder
Source: https://github.com/mashharuki/vibekanban-gitworktree-sample/tree/main/.claude/skills/mcp-builder
Command: npx skills add https://github.com/mashharuki/vibekanban-gitworktree-sample --skill mcp-builder-mashharuki

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: Guides deep research of MCP specs and SDKs, implementation with Zod/Pydantic schemas, code review, and evaluation creation. - Language-Specific References: Includes detailed TypeScript (MCP SDK) and Python (FastMCP) implementation guides with project structure, tool registration patterns, and quality checklists. - Evaluation Harness: Ships Python scripts to run XML-based QA evaluations against MCP servers over stdio, SSE, or streamable HTTP transports. - Use Case: When asked to build an MCP server integrating an external API (e.g., a weather or GitHub service), this Skill walks through tool naming, input/output schemas, error handling, pagination, and generating 10 verifiable evaluation questions. ## Quick Start Ask the AI to build an MCP server for your chosen API using the mcp-builder skill, specifying TypeScript or Python and the target transport.

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 input schemas, explicit descriptions, and annotations. The Skill's node_mcp_server.md reference covers project structure, tsconfig setup, and complete working examples for stdio and streamable HTTP transports.

How do I test and evaluate an MCP server?▼

Create an XML file with 10 read-only, verifiable QA pairs, then run scripts/evaluation.py with your transport type (stdio, sse, or http). The harness uses Claude to answer questions via your tools and reports accuracy, tool call counts, and per-task feedback.

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

Use stdio for local integrations and command-line tools where the server runs as a subprocess. Use streamable HTTP for remote servers, multi-client scenarios, and cloud deployments; SSE is deprecated in favor of streamable HTTP.

What makes a good MCP tool design?▼

Tools need snake_case names with service prefixes (e.g., github_create_issue), Zod or Pydantic input validation, concise descriptions, and annotations like readOnlyHint and destructiveHint. Responses should support JSON and Markdown formats with pagination metadata for list operations.

What dependencies are required to run the MCP evaluation scripts?▼

The evaluation harness requires the anthropic and mcp Python packages, plus an ANTHROPIC_API_KEY environment variable. It connects to your server via stdio, SSE, or HTTP and runs the agent loop against your evaluation XML file.