Claude Manager Orchestrator MCP Server

An MCP (Model Context Protocol) server that allows Claude to manage other Claude agents in the Claude Manager app.

Installation

cd mcp-orchestrator
npm install
npm run build

Configuration

Add to your Claude Code MCP settings (~/.claude/mcp.json):

{
  "mcpServers": {
    "claude-mgr-orchestrator": {
      "command": "node",
      "args": ["/path/to/dorothy/mcp-orchestrator/dist/index.js"]
    }
  }
}

Or for development:

{
  "mcpServers": {
    "claude-mgr-orchestrator": {
      "command": "npx",
      "args": ["tsx", "/path/to/dorothy/mcp-orchestrator/src/index.ts"]
    }
  }
}

Requirements

The Claude Manager Electron app must be running for the MCP server to work. It exposes a local API on port 31415.

Available Tools

list_agents

List all agents and their current status.

get_agent

Get detailed information about a specific agent.

• id: The agent ID

get_agent_output

Get the recent terminal output from an agent.

• id: The agent ID
• lines: Number of lines to retrieve (default: 100)

create_agent

Create a new agent for a specific project.

• projectPath: Absolute path to the project directory
• name: Name for the agent (optional)
• skills: List of skill names to enable (optional)
• character: Visual character (optional)
• skipPermissions: Run with --dangerously-skip-permissions (optional)
• secondaryProjectPath: Additional project context (optional)

start_agent

Start an agent with a specific task/prompt.

• id: The agent ID
• prompt: The task or instruction
• model: Optional model to use

stop_agent

Stop a running agent.

• id: The agent ID

send_message

Send input to an agent waiting for input.

• id: The agent ID
• message: The message to send

remove_agent

Permanently remove an agent.

• id: The agent ID

wait_for_agent

Poll an agent's status until it completes or errors.

• id: The agent ID
• timeoutSeconds: Maximum wait time (default: 300)
• pollIntervalSeconds: Poll interval (default: 5)

Example Usage

When configured, the Super Agent can:

"Create 3 agents for my project at /Users/me/myapp:
1. Backend agent to refactor the API
2. Frontend agent to update the UI
3. Test agent to write tests

Start them in sequence, waiting for each to complete before starting the next."

The orchestrator will use the tools to create, start, monitor, and coordinate the agents.