mcp-builder

name	mcp-builder
description	Use when building MCP servers or integrating external APIs/services via MCP (TypeScript or Python).
allowed-tools	["shell","read_file","write_file","apply_patch","update_plan"]
metadata	{"short-description":"Build MCP servers and tools"}

MCP Server Development Guide

Overview

Create MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. The quality of an MCP server is measured by how well it enables real-world tasks.

---

Process

Phase 1: Deep Research and Planning

1.1 Understand Modern MCP Design

API Coverage vs. Workflow Tools: Balance comprehensive endpoint coverage with specialized workflow tools. When uncertain, prefer comprehensive coverage.

Tool Naming and Discoverability: Use clear, consistent, action-oriented tool names.

Context Management: Keep tool descriptions concise; support filtering/pagination.

Actionable Error Messages: Errors should guide agents toward resolution with specific next steps.

1.2 Study MCP Protocol Documentation

Start with the sitemap and fetch markdown pages using web.run:

• https://modelcontextprotocol.io/sitemap.xml
• https://modelcontextprotocol.io/specification/draft.md

Review architecture, transports, tool/resource/prompt definitions.

1.3 Study Framework Documentation

Recommended stack:

• Language: TypeScript
• Transport: Streamable HTTP for remote servers; stdio for local servers

Load framework documentation:

• MCP Best Practices: reference/mcp_best_practices.md
• TypeScript SDK: https://raw.githubusercontent.com/modelcontextprotocol/typescript-sdk/main/README.md
• TypeScript Guide: reference/node_mcp_server.md
• Python SDK: https://raw.githubusercontent.com/modelcontextprotocol/python-sdk/main/README.md
• Python Guide: reference/python_mcp_server.md

1.4 Plan Your Implementation

• Review API docs to identify endpoints, auth, and data models
• Prioritize common operations and ensure pagination support

---

Phase 2: Implementation

2.1 Set Up Project Structure

See language-specific guides in reference/.

2.2 Implement Core Infrastructure

• Authenticated API client
• Error handling helpers
• Response formatting
• Pagination utilities

2.3 Implement Tools

Input Schema: Zod (TS) or Pydantic (Python) with constraints and examples

Output Schema: Define outputSchema where possible; return structured content

Annotations:

• readOnlyHint, destructiveHint, idempotentHint, openWorldHint

---

Phase 3: Review and Test

3.1 Code Quality

• DRY, consistent error handling, type coverage, clear descriptions

3.2 Build and Test

TypeScript:

• npm run build
• npx @modelcontextprotocol/inspector

Python:

• python -m py_compile your_server.py
• MCP Inspector

---

Phase 4: Create Evaluations

Load reference/evaluation.md and create 10 realistic, stable, read-only evaluation questions with verifiable answers.

---

Reference Files

These files are bundled with this skill under the local reference/ directory:

• reference/mcp_best_practices.md
• reference/node_mcp_server.md
• reference/python_mcp_server.md
• reference/evaluation.md

Arguments

Target: ${ARGUMENTS}