mcp-integration

Configure MCP servers in Claude Code plugins via .mcp.json.

15|4|Updated Apr 18, 2026

Search Tags:#automation #authentication #integration #mcp #plugin #claude-code #server-types

Ready to run

npx skills add https://github.com/damionrashford/media-os --skill mcp-integration-damionrashford

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: mcp-integration
Source: https://github.com/damionrashford/media-os/tree/main/.claude/plugins/plugin-dev/skills/mcp-integration
Command: npx skills add https://github.com/damionrashford/media-os --skill mcp-integration-damionrashford

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

MCP integration enables Claude Code plugins to securely expose external services as tools via the Model Context Protocol, allowing plugins to interact with external tools, databases, and APIs without bespoke wiring, while supporting multiple server types and secure configurations.

Core Features & Use Cases

Expose external services as MCP tools within Claude Code plugins via .mcp.json or plugin.json configurations.
Support for multiple MCP server types (stdio, SSE, HTTP, WebSocket) with either OAuth or token-based authentication.
Environment-variable expansion, automatic credential handling, and lifecycle management across server types.
Suitable for multi-server plugin architectures, enterprise integrations, and automated tool orchestration across services.

Quick Start

Create a .mcp.json at your plugin root describing your MCP server, then reference tools using the mcp__plugin_<plugin>_<server>__<tool> names in commands.

Frequently Asked Questions about mcp-integration

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

FAQPage Schema

How do I integrate MCP servers into Claude Code plugins?▼

To integrate MCP servers, create a .mcp.json file at your plugin root or add an mcpServers entry to plugin.json. You then reference tools using the mcp__plugin_<plugin>_<server>__<tool> naming convention within your commands.

What MCP server types are supported for Claude Code plugin integration?▼

MCP integration supports four server types: local stdio, hosted SSE, HTTP, and WebSocket servers. Each type handles secure configurations, environment-variable expansion, and automatic credential management across the server lifecycle.

How does authentication work when exposing external services as MCP tools?▼

Authentication for MCP tools uses either OAuth or token-based authentication depending on the server type. The integration handles automatic credential passing and environment-variable expansion to securely connect external services without bespoke wiring.

Can I use multiple MCP servers in a single Claude Code plugin architecture?▼

Yes, MCP integration supports multi-server plugin architectures and enterprise setups. You can orchestrate automated tools across multiple services simultaneously by configuring multiple server definitions within your .mcp.json or plugin.json file.

Why do I need to prefix MCP tool names with mcp__ in Claude Code?▼

The mcp__ prefix is required for tool naming so Claude Code can correctly route commands to the appropriate plugin and server. The full format is mcp__plugin_<plugin>_<server>__<tool>, ensuring unique identification across multi-server architectures.

Does MCP integration require documenting environment variables and scopes?▼

Yes, configuring MCP integration requires documenting required environment variables and scopes for secure plugin operation. This ensures proper credential handling and lifecycle management across the chosen stdio, SSE, HTTP, or WebSocket server types.