Can I use different AI models with Forky?

Yes, you can specify models using the -m flag, allowing you to use lighter models like Haiku for simple tasks and more powerful models like Opus for complex design work.

What is the Forky skill for Claude Code?

Forky is a skill that allows you to spawn multiple parallel Claude sessions to handle different tasks simultaneously without losing your current working context.

Can I interact with a fork after it has started?

Absolutely. You can use the -l flag or specific IDs to send follow-up messages to an existing fork if you need it to refine its work.

How do I know when a background fork is finished?

Forks run detached, and you will receive a notification via a Stop hook once the fork calls the 'done' command with its results summary.

Does Forky preserve my project context?

Yes, each fork is initialized with your current context, ensuring the background session understands your code and objectives as well as the main session does.

Forky - Parallel Claude Sessions

Name: Forky - Parallel Claude Sessions
Author: tomWhiting

bytomWhiting

0•

Productivity & Workflow

Parallelizes Claude tasks by spawning detached background sessions to handle research, testing, and documentation without interrupting your primary workflow.

Forky is a specialized productivity skill for Claude Code designed to solve the bottleneck of linear task execution. It enables you to delegate 'side quests'—such as running extensive test suites, researching API documentation, or generating boilerplate—to independent, asynchronous Claude sessions. By offloading these time-consuming or distracting tasks to background 'forks,' you can maintain focus on your main objective while the system notifies you upon the completion of each sub-task, effectively turning Claude into a multi-threaded development assistant.

Key Features

01Automatic completion notifications via system Stop hooks

02Spawn detached background Claude sessions for parallel execution

03Session continuation support to message existing forks without starting over

040 GitHub stars

05Manage and track multiple concurrent 'tines' (forks) with unique IDs

06Custom model selection (Haiku/Opus) per fork to optimize for speed or complexity

Use Cases

01Researching multiple architectural approaches or documentation sets simultaneously

02Running long-running test suites or linting processes while continuing to write new code

03Delegating housekeeping tasks like README updates or comment generation while focusing on core logic

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add tomwhiting/forky forky

For use in Claude.ai and ChatGPT

Download Skill