1.0.14 • Published 5 months ago

@hive-academy/mcp-workflow-manager v1.0.14

Weekly downloads
-
License
MIT
Repository
github
Last release
5 months ago

MCP Workflow Manager

A comprehensive Model Context Protocol (MCP) server for AI workflow automation and task management. Built with NestJS, Prisma, and @rekog/mcp-nest for production-ready reliability.

✨ Key Features

  • šŸ”„ Role-Based Workflow: Structured AI coordination (Boomerang, Researcher, Architect, Developer, Code Review)
  • šŸ“Š Task Management: Complete task lifecycle with status tracking and delegation
  • šŸŽÆ Implementation Planning: Batch-based subtask organization and execution
  • šŸ“ˆ Analytics & Reporting: Comprehensive workflow analytics and progress monitoring
  • šŸ”’ Project Isolation: Automatic database separation for multi-project workflows
  • šŸš€ Zero Setup: Just add to MCP config - no manual installation required
  • šŸ“¦ Self-Contained NPX Package: Automatic dependency management with no external requirements
  • šŸ”§ Environment-Aware: Adapts behavior for NPX, global, and local installations

šŸš€ Quick Setup Guide

Streamlined Database Configuration ✨

Zero Configuration Required! Each project automatically gets its own isolated database:

project-a/data/workflow.db  ← Project A's data
project-b/data/workflow.db  ← Project B's data
project-c/data/workflow.db  ← Project C's data

NPX Setup (Recommended)

One command, automatic project isolation:

{
  "mcpServers": {
    "workflow-manager": {
      "command": "npx",
      "args": ["-y", "@hive-academy/mcp-workflow-manager"]
    }
  }
}

āœ… Automatic project detection
āœ… Database created in ./data/workflow.db
āœ… Migrations applied safely
āœ… Zero setup required

Docker Setup

Project isolation via volume mounts:

{
  "mcpServers": {
    "workflow-manager": {
      "command": "docker",
      "args": [
        "run",
        "--rm",
        "-i",
        "-v",
        "D:/projects/your-project/data:/app/data",
        "hiveacademy/mcp-workflow-manager"
      ]
    }
  }
}

Replace D:/projects/your-project/ with your actual project path

āœ… Instant startup (pre-deployed migrations)
āœ… Project isolation via different volume mounts
āœ… Consistent across platforms

How Project Isolation Works

NPX: Automatic detection

cd /path/to/project-a && npx @hive-academy/mcp-workflow-manager
# Database: /path/to/project-a/data/workflow.db

cd /path/to/project-b && npx @hive-academy/mcp-workflow-manager
# Database: /path/to/project-b/data/workflow.db

Docker: Volume mount isolation

# Project A
docker run -v "/path/to/project-a/data:/app/data" hiveacademy/mcp-workflow-manager

# Project B
docker run -v "/path/to/project-b/data:/app/data" hiveacademy/mcp-workflow-manager

šŸ› ļø Available MCP Tools

Core Workflow Management

  • task_operations - Task lifecycle management (create, update, get, list)
  • planning_operations - Implementation planning and batch subtask management
  • workflow_operations - Role-based delegation and workflow transitions
  • review_operations - Code review and completion report management
  • research_operations - Research reports and communication management

Query & Analytics

  • query_task_context - Comprehensive task context retrieval
  • query_workflow_status - Delegation and workflow status queries
  • query_reports - Report queries with evidence relationships
  • batch_subtask_operations - Bulk subtask management by batch
  • batch_status_updates - Cross-entity status synchronization

Reporting & Analytics

  • generate_workflow_report - Comprehensive workflow analytics and reports
  • get_report_status - Report generation status tracking
  • cleanup_report - Report file management

šŸ”„ Workflow Roles

The system implements a structured workflow with specialized AI roles:

  1. 🪃 Boomerang - Task intake, analysis, and final delivery
  2. šŸ”¬ Researcher - Information gathering and research
  3. šŸ›ļø Architect - Technical planning and design
  4. šŸ‘Øā€šŸ’» Senior Developer - Code implementation
  5. šŸ” Code Review - Quality assurance and testing

āœ… Verification

After adding the configuration:

  1. Restart your MCP client
  2. Check for workflow tools in your MCP client
  3. Create a test task to verify everything works

You should see 10+ workflow management tools available!

šŸ“š Documentation

šŸ¤ Contributing

We welcome contributions! See our GitHub repository for details.

šŸ“„ License

MIT License - see LICENSE file.


Built with ā¤ļø for the AI development community by Hive Academy