How does Subtask prevent code conflicts between AI workers?

Subtask creates a unique git worktree for every task. This ensures each worker operates in an isolated directory with its own file state, preventing them from overwriting each other's work.

What happens when a task is finished?

Once a task is complete, you can review the changes using 'subtask diff'. After approval, the 'subtask merge' command squash-merges all changes into your base branch.

Can I monitor what an AI worker is thinking during a task?

Yes, you can use the 'subtask trace' command to debug the internal reasoning of a worker, or 'subtask log' to view the full conversation and event history.

Does Subtask support background tasks?

Yes. By using the Bash tool with background execution enabled, you can dispatch workers and receive notifications when they finish, allowing you to continue other work in the meantime.

Subtask Orchestrator

Name: Subtask Orchestrator
Author: zippoxer

byzippoxer

•

229

Collaboration Tools

Orchestrates parallel AI workers in isolated git worktrees to execute complex development tasks autonomously.

About

Subtask is a powerful CLI-driven orchestration skill for Claude Code that enables the delegation of software development tasks to multiple parallel AI subagents. By leveraging isolated git worktrees, it allows developers to dispatch numerous tasks simultaneously—such as bug fixes, feature development, or refactoring—without the risk of version control conflicts. This skill empowers a 'Lead' AI to clarify requirements, manage worker lifecycles, and review output, significantly increasing development throughput and streamlining complex project management.

Key Features

Comprehensive task lifecycle management from drafting to squash-merging.
Parallel task execution using isolated git worktrees to prevent code conflicts.
Real-time monitoring via task logs and internal reasoning traces.
Background execution support allowing for asynchronous task completion.
229 GitHub stars
Autonomous planning workflows with mutual review between lead and workers.

Use Cases

Managing large-scale refactoring projects by breaking them into manageable, reviewed sub-tasks.
Executing multiple independent bug fixes simultaneously across a large codebase.
Delegating routine feature implementations to subagents while maintaining architectural oversight.

About

Key Features

Comprehensive task lifecycle management from drafting to squash-merging.
Parallel task execution using isolated git worktrees to prevent code conflicts.
Real-time monitoring via task logs and internal reasoning traces.
Background execution support allowing for asynchronous task completion.
229 GitHub stars
Autonomous planning workflows with mutual review between lead and workers.

Use Cases

Managing large-scale refactoring projects by breaking them into manageable, reviewed sub-tasks.
Executing multiple independent bug fixes simultaneously across a large codebase.
Delegating routine feature implementations to subagents while maintaining architectural oversight.