mcp-integration

Configure MCP servers and expose tools within Claude Code plugins.

1.0k|91|Updated Jul 9, 2025
One-click install
npx skills add https://github.com/fcakyon/claude-codex-settings --skill mcp-integration-fcakyon
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: mcp-integration
Source: https://github.com/fcakyon/claude-codex-settings/tree/main/plugins/plugin-dev/skills/mcp-integration
Command: npx skills add https://github.com/fcakyon/claude-codex-settings --skill mcp-integration-fcakyon

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes examples (resource) and references (resource) components.

What problem does it solve?

Integrating Claude Code plugins with diverse external services and APIs can be challenging, requiring specific knowledge of protocols and authentication. This Skill simplifies the process, enabling rapid connection to any service, saving development effort and expanding Claude's capabilities.

Core Features & Use Cases

  • Flexible Server Configuration: Configure Model Context Protocol (MCP) servers using .mcp.json or inline in plugin.json.
  • Diverse Server Types: Integrate with stdio (local processes), SSE (hosted cloud services with OAuth), HTTP (REST APIs), and WebSocket (real-time streaming) servers.
  • Automated Authentication: Claude Code handles OAuth flows automatically, simplifying secure connections to cloud services.
  • Portable Environment Variables: Use ${CLAUDE_PLUGIN_ROOT} and user environment variables for flexible and secure configuration.
  • Use Case: Connect your plugin to a project management tool like Asana, integrate with a custom internal REST API, or expose a local database as a Claude tool for automated data operations.

Quick Start

Show me how to configure an MCP server in my plugin to connect to a REST API using a bearer token.

Frequently Asked Questions about mcp-integration

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

FAQPage Schema
How do I connect Claude to external APIs and services using the Model Context Protocol?

The Model Context Protocol (MCP) enables Claude Code plugins to integrate with external services through stdio, SSE, HTTP, or WebSocket servers. Configure MCP servers in `.mcp.json` or inline in `plugin.json`, expose tools with the naming convention `mcp__plugin_<plugin-name>_<server-name>__<tool-name>`, and Claude handles authentication and lifecycle management automatically.

Can I use OAuth authentication with MCP server integration in Claude plugins?

Yes, MCP integration supports automated OAuth flows for secure cloud service connections. Claude Code handles the OAuth authentication process automatically, eliminating manual token management while maintaining security best practices across local and hosted deployments.

What server types does MCP integration support for connecting to external services?

MCP integration supports four server types: stdio for local processes, SSE for hosted cloud services, HTTP for REST APIs, and WebSocket for real-time streaming connections. Each type accommodates different service architectures and communication patterns.

How do I configure environment variables and API credentials for MCP servers in plugins?

MCP integration uses portable environment variables like `${CLAUDE_PLUGIN_ROOT}` and user environment variables for flexible configuration. Dynamic headers and credentials are managed securely across deployment contexts through `.mcp.json` or `plugin.json` configuration with automatic environment variable expansion.

Can I connect Claude plugins to private REST APIs and internal services?

Yes, MCP integration enables connection to custom internal REST APIs and private services using bearer token authentication and HTTP server configuration. Configure token-based authentication in your plugin's MCP settings for secure access to proprietary endpoints.

What naming convention must MCP tools follow in Claude Code plugins?

MCP tools must follow the standardized naming convention `mcp__plugin_<plugin-name>_<server-name>__<tool-name>`. This enforced naming scheme ensures consistent tool identification, proper scoping across multiple MCP servers, and reliable integration within the Claude Code plugin ecosystem.