MCP Integration Overview

Station is primarily an MCP runtime that augments your Claude Code and Cursor development experience by providing an agnostic runtime for your sub-agents.

Station as MCP Runtime

Rather than managing agents in various platforms, Station gives you:

  • Centralized Agent Management - One place to create, manage, and monitor all your sub-agents
  • Environment Isolation - Separate agents and MCP tools by dev/staging/production environments
  • Server Deployment - Deploy Station as a persistent server for background agent operations
  • Universal MCP Client - Works with any MCP-compatible tools and servers
graph LR
    subgraph "Development Environment"
        Claude[Claude Code]
        Cursor[Cursor]
        Station[Station Runtime]
    end
    
    subgraph "Sub-Agents"
        Agent1["File Monitor<br/>📁 list_directory<br/>📄 read_text_file"]
        Agent2["Code Reviewer<br/>🔍 search_files<br/>📝 edit_file"]
        Agent3["Deploy Agent<br/>🐳 docker_exec<br/>☸️ kubectl_apply"]
    end
    
    subgraph "Dev Environment MCP Pool"
        FS[Filesystem MCP]
        Docker[Docker MCP]
        K8s[Kubernetes MCP]
        DB[Database MCP]
        Custom[Custom Tools]
    end
    
    Claude --> Station
    Cursor --> Station
    Station --> Agent1
    Station --> Agent2
    Station --> Agent3
    
    Agent1 --> FS
    Agent2 --> FS
    Agent3 --> Docker
    Agent3 --> K8s
    Agent1 --> DB

Key Benefits:

  • Tool Reuse - Share MCP servers across all your development tools
  • Agent Portability - Export agents as .prompt files for version control
  • Background Execution - Long-running agents that work while you code
  • Multi-Environment - Same agent, different tools per environment

MCP Templates & Variable Resolution

Station treats MCP configurations as templates with environment-specific variable resolution. This allows you to define MCP servers once and deploy them across multiple environments with different settings.

Template System

MCP configurations support Go template syntax for dynamic values:

{
  "name": "Filesystem MCP",
  "description": "Basic filesystem operations",
  "mcpServers": {
    "filesystem": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-filesystem@latest",
        "{{ .ALLOWED_PATH }}"
      ]
    }
  }
}

Variable Resolution with stn sync

When you run stn sync, Station automatically resolves template variables:

stn sync default

Station will:

  1. Look for variables.yml in your environment directory
  2. Prompt for missing variables if not found in the file
  3. Resolve all templates using the variables
  4. Discover tools from resolved MCP servers

Quick Start

Get Station running as an MCP server in Claude Desktop in 5 minutes:

1. Install & Initialize

curl -fsSL https://raw.githubusercontent.com/cloudshipai/station/main/install.sh | bash
stn init

2. Load MCP Tools

# Create filesystem MCP template
cat > filesystem-mcp.json << 'EOF'
{
  "name": "Filesystem MCP",
  "description": "Basic filesystem operations", 
  "mcpServers": {
    "filesystem": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-filesystem@latest", "{{ .ALLOWED_PATH }}"]
    }
  }
}
EOF

cp filesystem-mcp.json ~/.config/station/environments/default/template.json
echo "ALLOWED_PATH: /home/user/projects" > ~/.config/station/environments/default/variables.yml
stn sync

3. Connect to Claude Desktop

Add to your Claude Desktop config:

{
  "mcpServers": {
    "station": {
      "command": "stn",
      "args": ["stdio"]
    }
  }
}

4. Create Agents

You: “Create an agent that monitors project files”

Station automatically assigns the right filesystem tools and creates your deployable sub-agent!

Next Steps