fan-out

Run multiple Claude agents in parallel across isolated git worktrees.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/vr000m/skills.md --skill fan-out-vr000m
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: fan-out
Source: https://github.com/vr000m/skills.md/tree/main/.claude/skills/fan-out
Command: npx skills add https://github.com/vr000m/skills.md --skill fan-out-vr000m

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Fans out independent tasks from a dev-plan to parallel Claude agents running in isolated git worktrees, then merges and verifies the results. Use after planning when 2+ tasks can run independently, or when the user says "fan out", "parallelize this", or "/fan-out".

Core Features & Use Cases

  • Dispatch independent tasks to parallel Claude agents, each in its own worktree
  • Merge and verify results across branches to ensure coherence
  • One-level fan-out with deterministic worktree and branch naming to simplify rollback and auditing

Quick Start

Spawn parallel agents for a ready dev plan to execute independent tasks and merge results.

Frequently Asked Questions about fan-out

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run parallel agents in isolated git worktrees to execute independent tasks?

You can run parallel agents in isolated git worktrees by using a fan-out orchestration process that spawns per-task worktrees and branches, dispatches the tasks to agents, and then merges and verifies the results.

What is a one-level fan-out boundary in parallel agent orchestration?

A one-level fan-out boundary restricts parallel agent orchestration to a single layer of task dispatch, preventing recursive spawning to simplify rollback, monitoring, and state auditing across isolated git worktrees.

How do I parallelize a dev plan when multiple tasks can run independently?

To parallelize a dev plan, you dispatch independent tasks to parallel Claude agents, each operating in its own isolated git worktree, ensuring deterministic branch naming before merging the verified results.

Can I use git worktrees to isolate parallel Claude agents for independent coding tasks?

Yes, you can isolate parallel Claude agents by assigning each agent its own git worktree and branch, providing per-task context to execute independent coding tasks before merging the results.

What are the limitations of using a one-level fan-out for parallel agent orchestration?

The limitation of one-level fan-out orchestration is that it restricts parallel agent execution to a single dispatch layer, preventing deeper recursive task spawning while ensuring reproducible workflows and simplified auditing.

When should I parallelize independent tasks across isolated git worktrees?

You should parallelize independent tasks across isolated git worktrees after planning when you have two or more tasks that can run independently and need concurrent execution with reproducible merging and verification.