mcp-builder
mcp-builder covers the essentials of building Model Context Protocol servers, from project structure and transport selection to tool and resource design. It walks through input validation, error handling, multimodal support, and configuration patterns for both local and remote deployments. Follow the included checklist to ensure your server is well-documented, secure, and ready for production use.
mcp-builder teaches you to construct MCP servers with sound architecture, tool patterns, and security practices.
AI-generated summary based on this skill's SKILL.md
Install
vudovn/ag-kit/mcp-builder · repository language: TypeScript
git clone https://github.com/vudovn/ag-kit
cp -r ag-kit/.agents/skills/mcp-builder ~/.claude/skills/mcp-buildernpx skillfed install vudovn/ag-kit/mcp-builderFrequently asked questions
AI-generated answers based on this skill's SKILL.md and metadata
How do I build MCP servers with proper architecture?
mcp-builder teaches you to construct Model Context Protocol servers by starting with project structure, selecting the right transport (stdio for local, HTTP for remote), and designing tools and resources following best practices. The skill covers tool naming conventions, input schema validation, error handling, and multimodal content support. It includes a production-readiness checklist to ensure your server is well-documented, secure, and properly configured for deployment.
What are MCP tool design best practices?
mcp-builder emphasizes designing MCP tools by defining clear input schemas with proper validation, using consistent naming conventions, and structuring resources with meaningful URI patterns. Best practices include implementing robust error handling, supporting multimodal content where applicable, and documenting tool behavior thoroughly. The skill guides you through patterns that make tools predictable and maintainable for AI clients like Claude.
How do I configure MCP servers for Claude Desktop?
mcp-builder covers MCP server configuration for Claude Desktop by explaining transport setup, authentication patterns including OAuth, and local vs. remote deployment options. You'll learn how to structure your server configuration, validate inputs, and handle errors gracefully. The skill includes practical guidance on testing strategies and a checklist to verify your server meets production requirements before connecting it to Claude Desktop.
What should I know about MCP input schema validation?
mcp-builder teaches MCP input schema validation as a core security and reliability practice. You'll learn to define structured schemas that validate tool parameters, implement proper error handling for invalid inputs, and follow validation patterns that prevent malformed requests. The skill emphasizes validation's role in maintaining server stability and ensuring AI clients receive clear feedback when inputs don't meet requirements.
How do I implement MCP security and error handling?
mcp-builder covers MCP security principles including input validation, proper error handling, and authentication setup (OAuth patterns included). The skill teaches you to design servers that fail safely, return meaningful error messages, and protect against common vulnerabilities. Error handling guidance includes distinguishing between client errors and server issues, logging appropriately, and ensuring your server remains stable under unexpected conditions.
What transport options exist for Model Context Protocol servers?
mcp-builder explains MCP transport choices: stdio for local server-to-client communication (ideal for Claude Desktop integration) and HTTP for remote deployments. The skill helps you select the right transport based on your architecture needs, configure it correctly, and understand the trade-offs. Both transports are covered with practical setup guidance and best practices for production deployment.
SKILL.md
rendered from the published skill — quoted content, verbatim
MCP Builder
> Principles for building MCP servers.
1. MCP Overview
What is MCP?
Model Context Protocol - standard for connecting AI systems with external tools and data sources.
Core Concepts
| Concept | Purpose |
|---|---|
| Tools | Functions AI can call |
| Resources | Data AI can read |
| Prompts | Pre-defined prompt templates |
2. Server Architecture
Project Structure
my-mcp-server/
├── src/
│ └── index.ts # Main entry — McpServer from @modelcontextprotocol/sdk
├── package.json # dep: @modelcontextprotocol/sdk
└── tsconfig.json
> TypeScript SDK: @modelcontextprotocol/sdk (McpServer). Python SDK: mcp (pip install "mcp[cli]", FastMCP bundled in mcp.server.fastmcp).
Transport Types
| Type | Use |
|---|---|
| stdio | Local, CLI-based |
(truncated - see the full file via the links below)
Read as markdown · JSON record · Browse the source repository
File tree — 1 file
.agents/skills/mcp-builder/SKILL.md