mcp-config

Automate MCP server configuration across project and user scopes.

1|Updated Nov 15, 2025
One-click install
npx skills add https://github.com/racurry/neat-little-package --skill mcp-config
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: mcp-config
Source: https://github.com/racurry/neat-little-package/tree/main/plugins/box-factory/skills/mcp-config
Command: npx skills add https://github.com/racurry/neat-little-package --skill mcp-config

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Configuring MCP servers involves transport choices, scopes, and security patterns. This skill provides guidance to help you decide how to configure MCP servers across projects, users, and plugins.

Core Features & Use Cases

  • Transport guidance: When to choose HTTP versus stdio transports
  • Scope guidance: Project vs user vs local scope guidance
  • Security patterns: Best practices for storing secrets and tokens

Quick Start

Example user request: Recommend an MCP transport and scope for a new server and outline the security considerations.

Frequently Asked Questions about mcp-config

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I choose between HTTP and stdio transports for MCP servers?

HTTP and stdio transports serve different deployment patterns. HTTP works best for remote or distributed servers, while stdio is simpler for local integrations. Choose HTTP for network isolation and scalability; use stdio for tightly coupled local processes. Security, latency, and your infrastructure determine the fit.

Where should I store MCP server configurations—project-level or user-level?

Project-level configuration in .mcp.json keeps server setup with your codebase for team sharing; user-level in ~/.claude.json applies across all projects. Use project scope for shared integrations and user scope for personal tools or credentials you don't commit.

How do I securely handle secrets and tokens in MCP server configuration?

Store secrets in environment variables and reference them in .mcp.json or ~/.claude.json, never hardcode credentials. Use scoped environment variables per server, document required keys, and validate that all secrets are set before initialization.

What's the process for adding and validating new MCP servers in Claude Code?

Add server definitions to the appropriate scope configuration file, specify transport and credentials, then validate using built-in listing and health-check workflows. Document each server's purpose, transport choice, and required environment variables for reproducibility.

Can I configure MCP servers for both local development and production in the same project?

Yes, use environment-based configuration to switch servers between scopes. Define development servers locally or project-scoped, production servers user-scoped or with environment flags. This separation prevents accidental production use during development.