TaskFlow

TaskFlow MCP 🔄✅

A task management Model Context Protocol (MCP) server for planning and executing tasks with AI assistants.

🌟 Overview

TaskFlow MCP is a specialized server that helps AI assistants break down user requests into manageable tasks and track their completion. It enforces a structured workflow with user approval steps to ensure tasks are properly tracked and users maintain control over the process.

✨ Features

• 📋 Task Planning: Break down complex requests into manageable tasks
• 🔍 Subtasks: Divide tasks into smaller, more manageable subtasks
• 📊 Progress Tracking: Track the status of tasks, subtasks, and requests with visual progress tables
• 👍 User Approval: Enforce user approval steps to ensure quality and control
• 💾 Persistence: Save tasks and requests to disk for persistence across sessions
• 🔄 Flexible Management: Add, update, or delete tasks and subtasks as needed
• 📝 Detailed Reporting: View task details and progress tables
• 📤 Export Options: Export task plans and status reports in Markdown, JSON, or HTML formats
• 📦 Dependencies: Track project and task-level dependencies with version information
• 📌 Notes: Add project-level notes for important information and preferences
• 📄 YAML Support: Save tasks in YAML format for better handling of multiline content
• 🛡️ Robust Text Handling: Comprehensive newline sanitization for reliable data persistence
• 🎯 Prompts System: Global instructions and task prefix/suffix for consistent LLM guidance
• 📚 Task Archiving: Archive completed requests to keep active task lists clean
• 🗂️ Archive Management: Browse, search, and restore archived tasks with full history
• 📍 Relative Path Support: Use relative paths for flexible project-based workflows

🚀 Installation

Global Installation

npm install -g @pinkpixel/taskflow-mcp

Local Installation

npm install @pinkpixel/taskflow-mcp

🛠️ Usage

Starting the Server

If installed globally:

taskflow-mcp

If installed locally:

npx taskflow-mcp

Configuration

By default, TaskFlow MCP saves tasks to tasks.yaml in the current working directory. You can customize this by setting the TASK_MANAGER_FILE_PATH environment variable:

File Path Options

Absolute paths (recommended for production):

TASK_MANAGER_FILE_PATH=/home/user/projects/my-tasks.yaml taskflow-mcp
# Windows
TASK_MANAGER_FILE_PATH=C:\Users\username\Documents\tasks.yaml taskflow-mcp

Relative paths (great for project-based workflows):

# Resolves to ./project-tasks.yaml in the current directory
TASK_MANAGER_FILE_PATH=project-tasks.yaml taskflow-mcp

# Resolves to ./tasks/current.yaml relative to working directory
TASK_MANAGER_FILE_PATH=tasks/current.yaml taskflow-mcp

Advanced: Custom base directory

# Use a different base directory for relative path resolution
TASK_MANAGER_BASE_DIR=/home/user/workspace TASK_MANAGER_FILE_PATH=tasks.yaml taskflow-mcp

Cross-Platform Compatibility

TaskFlow MCP automatically handles path resolution across Windows and Linux:

• Uses Node.js path.resolve() and path.normalize() for consistent behavior
• Supports both forward slashes (/) and backslashes (\) on Windows
• Automatically creates parent directories when saving tasks
• Provides clear error messages for path resolution issues

YAML Format Support

TaskFlow MCP supports both JSON and YAML formats for data persistence. To use YAML format, simply configure your file path with a .yaml or .yml extension:

TASK_MANAGER_FILE_PATH=/path/to/tasks.yaml taskflow-mcp

YAML format is particularly useful for:

• Better preservation of multiline descriptions and text content
• More human-readable task data files
• Easier manual editing if needed

The format is automatically detected based on the file extension, and the system maintains full backward compatibility with existing JSON files.

Archive System

TaskFlow MCP v1.4.1 includes a comprehensive archive system to keep your active task lists clean while preserving completed work history:

# Configure archive file path (optional - defaults to [taskfile-name]-archive.[ext])
ARCHIVE_FILE_PATH=/path/to/tasks-archive.yaml taskflow-mcp

# Set archive mode (optional - defaults to 'manual')
ARCHIVE_MODE=manual taskflow-mcp  # or 'auto-on-complete'

Archive features include:

• Manual archiving: Use archive_completed_requests tool to archive when ready
• Automatic archiving: Set ARCHIVE_MODE=auto-on-complete for automatic archiving
• Archive browsing: Search and filter archived requests with list_archived_requests
• Archive restoration: Restore archived requests back to active status with restore_archived_request
• Full history preservation: Complete task history, timestamps, and metadata preserved

MCP Configuration

To use TaskFlow MCP with AI assistants, you need to configure your MCP client to use the server. Create an mcp_config.json file with the following content:

Basic Configuration:

{
  "mcpServers": {
    "taskflow": {
      "command": "npx",
      "args": ["-y", "@pinkpixel/taskflow-mcp"],
      "env": {
        "TASK_MANAGER_FILE_PATH": "/path/to/tasks.yaml"
      }
    }
  }
}

Advanced Configuration (with all v1.4.1 options):

{
  "mcpServers": {
    "taskflow": {
      "command": "npx",
      "args": ["-y", "@pinkpixel/taskflow-mcp"],
      "env": {
        "TASK_MANAGER_FILE_PATH": "./project-tasks.yaml",
        "TASK_MANAGER_BASE_DIR": "/path/to/project/root",
        "ARCHIVE_FILE_PATH": "./tasks-archive.yaml",
        "ARCHIVE_MODE": "manual"
      }
    }
  }
}

Configuration Options:

• TASK_MANAGER_FILE_PATH: Path to tasks file (supports .json/.yaml, absolute/relative paths)
• TASK_MANAGER_BASE_DIR: Custom base directory for relative path resolution
• ARCHIVE_FILE_PATH: Path to archive file (optional, auto-generated if not specified)
• ARCHIVE_MODE: Archive mode - "manual" (default) or "auto-on-complete"

💡 Tip: See examples/mcp_config_comprehensive.json for a complete configuration example with detailed comments and usage examples.

🔄 Workflow

TaskFlow MCP enforces a specific workflow:

1. Plan Tasks: Break down a user request into tasks (with optional subtasks)
2. Get Next Task: Retrieve the next pending task
3. Complete Subtasks: If the task has subtasks, complete each subtask before marking the task as done
4. Mark Task Done: Mark a task as completed (requires all subtasks to be completed first)
5. Wait for User Confirmation: Ask the user to confirm the completed task before proceeding
6. Repeat: Continue with the next task until all tasks are complete
7. Final Confirmation: Confirm with the user that the entire request has been completed

For AI assistants to consistently follow this workflow, see the example-system-prompt.md file for system prompts you can add to your assistant's instructions.

🧰 Available Tools

TaskFlow MCP exposes the following tools to AI assistants:

plan_task

Register a new user request and plan its associated tasks (with optional subtasks).

{
  "originalRequest": "Create a new website for my business",
  "outputPath": "C:/Users/username/Documents/website-project-plan.md",
  "dependencies": [
    {
      "name": "Node.js",
      "version": ">=14.0.0",
      "description": "JavaScript runtime"
    },
    {
      "name": "npm",
      "version": ">=6.0.0",
      "description": "Package manager"
    }
  ],
  "notes": [
    {
      "title": "Package Manager Preference",
      "content": "User prefers pnpm over npm for package management."
    },
    {
      "title": "Design Guidelines",
      "content": "Follow the company's brand guidelines for colors and typography."
    }
  ],
  "tasks": [
    {
      "title": "Design homepage",
      "description": "Create a design for the homepage with logo, navigation, and hero section",
      "dependencies": [
        {
          "name": "Figma",
          "description": "Design tool"
        }
      ],
      "subtasks": [
        {
          "title": "Design logo",
          "description": "Create a logo that represents the business brand"
        },
        {
          "title": "Design navigation",
          "description": "Create a user-friendly navigation menu"
        }
      ]
    },
    {
      "title": "Implement HTML/CSS",
      "description": "Convert the design to HTML and CSS",
      "dependencies": [
        {
          "name": "HTML5",
          "description": "Markup language"
        },
        {
          "name": "CSS3",
          "description": "Styling language"
        }
      ]
    }
  ]
}

get_next_task

Retrieve the next pending task for a request.

{
  "requestId": "req-1"
}

mark_task_done

Mark a task as completed.

{
  "requestId": "req-1",
  "taskId": "task-1",
  "completedDetails": "Created a modern design with a clean layout"
}

open_task_details

Get details about a specific task.

{
  "taskId": "task-1"
}

list_requests

List all requests in the system.

{}

add_tasks_to_request

Add more tasks to an existing request.

{
  "req

---

*README truncated. [View full README on GitHub](https://github.com/pinkpixel-dev/taskflow-mcp).*