Task Orchestrator MCP Server

A sophisticated task orchestration system with role-based workflow management and Trello integration.

Overview

The Task Orchestrator MCP server provides a comprehensive task management system with:

• Role-based workflow management with 5 distinct roles (Orchestrator, Architect, Coder, Analyst, DevOps)
• Three-tier Trello integration (MCP Server, Direct API, Local Storage)
• Automatic task lifecycle management with status tracking
• Dependency management between tasks
• Persistent local storage with JSON backup
• Real-time status monitoring and reporting

Features

Role System

• Orchestrator: Creates and assigns tasks, manages workflow
• Architect: Handles system design and architecture tasks
• Coder: Implements features and code changes
• Analyst: Performs analysis and research tasks
• DevOps: Manages deployment and infrastructure tasks

Trello Integration Modes

1. MCP Server Mode (Priority 1): Uses MCP Trello server for integration
2. Direct API Mode (Priority 2): Direct Trello API integration
3. Local Storage Mode (Priority 3): Local JSON storage without Trello

Task Management

• Automatic task creation with unique IDs
• Status tracking (TODO, IN_PROGRESS, REVIEW, DONE, BLOCKED)
• Dependency validation
• Role assignment and transitions
• Completion tracking with notes

Components

Resources

The server implements a task storage system with:

• Custom task:// URI scheme for accessing individual tasks
• Each task resource has a name, description and JSON mimetype
• Automatic resource list updates when tasks change

Tools

The server provides comprehensive task management tools:

Task Management

• create_task: Creates a new task (Orchestrator only)
• assign_task: Assigns task to a specific role (Orchestrator only)
• complete_task: Completes a task and returns control to Orchestrator
• list_tasks: Lists tasks with optional status filtering

Role Management

• switch_role: Switches to a different role (Orchestrator only)
• return_to_orchestrator: Returns control to Orchestrator
• get_status: Shows current system status and statistics

Trello Integration

• sync_to_trello: Syncs all tasks to Trello board
• check_mcp_trello: Checks MCP Trello server availability
• export_tasks: Exports tasks to local JSON files

Configuration

Environment Variables

For Direct API Trello integration:

TRELLO_API_KEY=your_api_key
TRELLO_TOKEN=your_token
TRELLO_WORKING_BOARD_ID=your_board_id

MCP Server Configuration

Development/Unpublished Servers

On MacOS: ~/Library/Application\ Support/Claude/claude_desktop_config.json On Windows: %APPDATA%/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "task-orchectrator-mcp": {
      "command": "uv",
      "args": [
        "--directory",
        "C:\\Users\\xella\\PycharmProjects\\task-orchectrator-mcp",
        "run",
        "task-orchectrator-mcp"
      ],
      "env": {
        "TRELLO_API_KEY": "your_api_key",
        "TRELLO_TOKEN": "your_token",
        "TRELLO_WORKING_BOARD_ID": "your_board_id"
      }
    }
  }
}

Published Servers

{
  "mcpServers": {
    "task-orchectrator": {
      "command": "npx",
      "args": ["@daymanking990/task-orchectrator-mcp"],
      "env": {
        "TRELLO_API_KEY": "your_trello_api_key",
        "TRELLO_TOKEN": "your_trello_token",
        "TRELLO_WORKING_BOARD_ID": "your_board_id"
      }
    }
  }
}

Quickstart

1. Install Dependencies

# Install py-trello for direct API integration
uv add py-trello

2. Configure Trello (Optional)

Set up environment variables for Trello integration:

export TRELLO_API_KEY="your_api_key"
export TRELLO_TOKEN="your_token"
export TRELLO_WORKING_BOARD_ID="your_board_id"

3. Start Using

The system will automatically:

• Detect available Trello integration modes
• Initialize in the best available mode
• Provide fallback to local storage if needed

Usage Examples

Create and Assign a Task

{
  "title": "Implement user authentication",
  "description": "Add OAuth2 authentication to the application",
  "dependencies": [],
  "create_trello_card": true
}

Check System Status

{}

Sync to Trello

{}

Development

Building and Publishing

Quick Publish

# Test locally
npm pack
npm install -g ./daymanking990-task-orchectrator-mcp-0.2.0.tgz

# Publish to npm
npm publish --access public

Manual Process

1. Update version in package.json, pyproject.toml, and __init__.py

2. Test locally:

npm pack
npm install -g ./daymanking990-task-orchectrator-mcp-0.2.0.tgz
task-orchectrator-mcp

3. Publish to npm:

npm publish --access public

GitHub Actions

The package is automatically published to npm when you create a GitHub release. See Build and Publish Guide for details.

Debugging

Use the MCP Inspector for debugging:

npx @modelcontextprotocol/inspector uv --directory C:\Users\xella\PycharmProjects\task-orchectrator-mcp run task-orchectrator-mcp

Documentation

• Trello Integration Guide - Complete guide to Trello integration modes
• Role System Guide - Detailed role management documentation
• MCP Setup Instructions - MCP server configuration guide
• Build and Publish Guide - How to build and publish the package

Architecture

The system uses a three-tier architecture for Trello integration:

1. MCP Server Layer: Highest priority, uses MCP protocol for Trello operations
2. Direct API Layer: Fallback to direct Trello API calls
3. Local Storage Layer: Final fallback with JSON file persistence

All data is automatically synchronized between layers when possible, ensuring data integrity and availability.

License

This project is licensed under the MIT License - see the LICENSE file for details.