MCP — Model Context Protocol

Connect external tools and data sources via the Model Context Protocol. Clew Code supports stdio, SSE, and DirectConnect transports.

Overview

MCP (Model Context Protocol) is an open standard for connecting AI agents with external tools and data sources. Clew Code implements the full MCP specification as both a client and server.

Transport Types

Transport	Description
stdio	Local subprocess — spawns an MCP server as a child process and communicates via stdin/stdout
SSE	Remote servers with optional OAuth authentication via Server-Sent Events
DirectConnect	In-process plugin servers — runs MCP servers as plugins within the same process

Configuration

MCP servers are configured in .mcp.json at the project root:

{
  "mcpServers": {
    "filesystem": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-filesystem", "."]
    },
    "github": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-github"],
      "env": {
        "GITHUB_TOKEN": "ghp_..."
      }
    },
    "remote-api": {
      "url": "https://my-mcp-server.example.com/sse",
      "auth": {
        "type": "oauth2"
      }
    }
  }
}

MCP Commands

❯ /mcp list         # list connected MCP servers and their tools
❯ /mcp add          # add a new MCP server configuration
❯ /mcp remove <id>  # remove an MCP server

CLI MCP Tools

The built-in MCP tools allow the AI to interact with connected servers:

Tool	Description
ListMcpResourcesTool	List available resources from connected MCP servers
ReadMcpResourceTool	Read a specific resource from an MCP server

Example MCP Servers

• Filesystem — safe file access with path restrictions
• GitHub — repository management, PRs, issues
• Playwright — browser automation
• Slack — workspace integration
• PostgreSQL — database schema exploration and queries
• Sentry — error tracking and issue management

Browse the official MCP server repository for more.