MCP Integration

Configure Model Context Protocol servers for Claude Code plugins.

1|Updated May 4, 2026
One-click install
npx skills add https://github.com/open-session/os_design_system --skill mcp-integration-open-session
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: MCP Integration
Source: https://github.com/open-session/os_design_system/tree/main/.claude/plugins/plugin-dev/skills/mcp-integration
Command: npx skills add https://github.com/open-session/os_design_system --skill mcp-integration-open-session

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill removes the complexity of connecting Claude Code plugins to external services by providing a clear, secure path for configuring Model Context Protocol servers, authentication, and tool access.

Core Features & Use Cases

  • Server Configuration: Set up MCP servers as stdio, SSE, HTTP, or WebSocket endpoints depending on the service and deployment model.
  • Authentication Guidance: Handle OAuth, bearer tokens, API keys, custom headers, and environment-variable-based credentials safely.
  • Tool Integration: Expose external capabilities as usable Claude Code tools with proper naming, permissions, and command-level allowlists.
  • Use Case: A plugin developer can connect a hosted SaaS API, a local file or database server, or a real-time service and make those tools available inside commands and agents.

Quick Start

Ask the assistant to configure your plugin’s MCP server, choose the right transport, and produce the needed auth and tool-usage setup.

Frequently Asked Questions about MCP Integration

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

FAQPage Schema
How do I configure an MCP server for Claude Code plugins?

Configuring an MCP server for Claude Code involves selecting a transport like stdio, SSE, HTTP, or WebSocket, setting portable ${CLAUDE_PLUGIN_ROOT} paths, and defining correctly named mcp__plugin tools for external service access.

What is the best way to handle authentication for Model Context Protocol integrations?

Handling authentication for Model Context Protocol integrations requires securely managing OAuth, bearer tokens, API keys, or custom headers using environment variables to protect credentials during tool exposure.

Can I use SSE and WebSocket transports to connect external services to Claude Code?

Yes, you can use SSE and WebSocket transports alongside stdio and HTTP to connect external services to Claude Code, enabling real-time tool access and agent workflows through secure WSS or HTTPS endpoints.

How do I expose external API capabilities as usable Claude Code tools?

Exposing external API capabilities as Claude Code tools requires configuring Model Context Protocol servers with proper tool naming conventions, command-level allowlists, and permissions to ensure secure agent workflow integration.

Why does my Claude Code MCP integration require portable paths and HTTPS endpoints?

Claude Code MCP integrations require portable ${CLAUDE_PLUGIN_ROOT} paths and secure HTTPS or WSS endpoints to ensure reliable plugin deployment and safe credential handling across different environments.