This project provides a Python CLI tool for coordinating multi-agent AI workflows, addressing the need for structured orchestration in complex AI development. It targets engineers and researchers working with AI agent systems, offering distinct modes (Linear, DAG, Debate) to streamline tasks ranging from simple bug fixes to intricate feature development and collaborative decision-making, thereby enhancing development efficiency and coordination.

How It Works

The core of the project is a standalone Python script (task_manager.py) that manages AI agent tasks through shared JSON files. It offers three distinct coordination modes: Linear for sequential pipelines, DAG for dependency-based parallel execution, and Debate for multi-agent deliberation and cross-review. This approach allows users to select the most appropriate workflow structure for their specific needs, optimizing agent collaboration and task management.

Quick Start & Requirements

• Installation: Clone the repository; no pip install is required as it's a standalone script.

  git clone https://github.com/win4r/team-tasks.git
  

  Run commands using python3 team-tasks/scripts/task_manager.py --help. For OpenClaw integration, copy the team-tasks/ directory into your OpenClaw skills directory.
• Prerequisites: Python 3.12+ (stdlib only).
• Dependencies: None external, relying solely on Python's standard library.
• Data Storage: Project data is stored as JSON in /home/ubuntu/clawd/data/team-tasks/ (configurable via TEAM_TASKS_DIR environment variable).
• Links: Repository: https://github.com/win4r/team-tasks.git

Highlighted Details

• Linear Mode: Supports sequential pipelines with auto-advance functionality.
• DAG Mode: Enables parallel task dispatch based on dependency satisfaction, featuring automatic unblocking notifications and cycle detection during task addition.
• Debate Mode: Facilitates multi-agent deliberation by collecting initial positions, generating cross-review prompts, and synthesizing outcomes.
• Comprehensive CLI: Offers a rich set of commands for project initialization, task management, status tracking, and DAG visualization.

Maintenance & Community

No specific details regarding maintainers, community channels (e.g., Discord, Slack), or roadmap are provided in the README.

Licensing & Compatibility

• License: MIT.
• Compatibility: The MIT license is permissive, generally allowing for commercial use and integration within closed-source projects.

Limitations & Caveats

The tool strictly requires Python 3.12+. Its standalone nature means it's not installed via package managers but requires cloning. The README highlights common pitfalls, such as incorrect stage ID usage in linear mode, dependency ordering in DAG, and adding debaters after debate rounds have commenced. Default data storage paths are specified but can be overridden via an environment variable.