How does this skill prevent agents from working on the same task?

It implements a 'claim-before-processing' rule where agents immediately update a shared task's status to 'In Progress' with a unique ID in the tasks.md file, effectively locking the task from other agents.

How do I track the progress of my multi-agent system?

All progress is centralized in a tasks.md file, which tracks status markers, commit hashes for completed work, and detailed error reasons if a task fails.

What is the benefit of using Git worktrees with this skill?

Git worktrees provide complete file system isolation, allowing multiple agents to run different tasks in separate directories simultaneously without interfering with each other's branch state or code changes.

Can I use different AI models for different tasks within the same queue?

Yes, the skill includes a tag system (e.g., #opus or #sonnet) that allows the orchestrator to route specific tasks to different models based on the task complexity or requirement.

Task-Based Multi-Agent Orchestrator

Name: Task-Based Multi-Agent Orchestrator
Author: melodic-software

bymelodic-software

•

生産性とワークフロー

Coordinates parallel AI agent execution using shared task files and Git worktree isolation for scalable workflows.

概要

This skill provides a robust framework for managing multi-agent systems within Claude Code, enabling parallel execution of complex tasks. By utilizing a centralized tasks.md file as a coordination mechanism and Git worktrees for environment isolation, it prevents race conditions and allows multiple agents to claim, process, and report on tasks concurrently. It is ideal for developers looking to scale their AI workflows beyond single-agent patterns into sophisticated, automated task queues with built-in status tracking, failure handling, and model-specific routing.

主な機能

Shared task coordination via Markdown files with standardized status markers
Git worktree isolation to prevent file conflicts during parallel agent execution
Integrated cron-trigger pattern for continuous polling and task processing
Tag-based routing for specific models (Opus/Sonnet) or specialized workflows
Automated task claiming system to prevent race conditions (Pending, In Progress, Complete)
1 GitHub stars

ユースケース

Implementing complex 'plan-then-implement' workflows with isolated environments
Scaling large-scale refactoring projects across multiple concurrent agents
Building automated task queue systems for background code processing

概要

主な機能

Shared task coordination via Markdown files with standardized status markers
Git worktree isolation to prevent file conflicts during parallel agent execution
Integrated cron-trigger pattern for continuous polling and task processing
Tag-based routing for specific models (Opus/Sonnet) or specialized workflows
Automated task claiming system to prevent race conditions (Pending, In Progress, Complete)
1 GitHub stars

ユースケース

Implementing complex 'plan-then-implement' workflows with isolated environments
Scaling large-scale refactoring projects across multiple concurrent agents
Building automated task queue systems for background code processing