Sandbox
@smtg-ai/claude-squad

Terminal manager for Claude Code and other agents

Claude Squad runs multiple agents in isolated workspaces so you can work on several tasks in parallel. It uses tmux for separate terminal sessions and git worktrees for branch isolation, with a TUI for creating, attaching to, reviewing, and resuming sessions.

8,467 stars619 forksGoUpdated 28d ago
Who it's for

Builders who want to run several terminal agents side by side without mixing their work.

What it delivers

You can keep multiple agent tasks moving at once and review each one before merging or pushing.

What it does

Multiple agent sessions

Starts each task in its own tmux session so agents can run in parallel.

Isolated git worktrees

Gives each session its own worktree to avoid branch conflicts.

Task review and checkout

Lets you inspect diffs, checkpoint work, and pause or resume sessions.

Profile-based launch commands

Lets you define named programs like `claude`, `codex`, and `aider` in the config file.

Autoyes mode

Can auto-accept prompts for supported agents with the `--autoyes` flag.

Session management UI

Provides keyboard-driven navigation, attach/detach, delete, and help actions in one terminal window.

How to get it

  1. 1Run
    brew install claude-squad
    ln -s "$(brew --prefix)/bin/claude-squad" "$(brew --prefix)/bin/cs"
  2. 2Claude Squad can also be installed by running the following command
    curl -fsSL https://raw.githubusercontent.com/smtg-ai/claude-squad/main/install.sh | bash
  3. 3To use a custom name for the binary
    curl -fsSL https://raw.githubusercontent.com/smtg-ai/claude-squad/main/install.sh | bash -s -- --name <your-binary-name>
  4. 4Run the application with
    cs

README

Claude Squad CI GitHub Release

Claude Squad is a terminal app that manages multiple Claude Code, Codex, Gemini (and other local agents including Aider) in separate workspaces, allowing you to work on multiple tasks simultaneously.

Claude Squad Screenshot

Highlights

  • Complete tasks in the background (including yolo / auto-accept mode!)
  • Manage instances and tasks in one terminal window
  • Review changes before applying them, checkout changes before pushing them
  • Each task gets its own isolated git workspace, so no conflicts

https://github.com/user-attachments/assets/aef18253-e58f-4525-9032-f5a3d66c975a


Installation

Both Homebrew and manual installation will install Claude Squad as cs on your system.

Homebrew

brew install claude-squad
ln -s "$(brew --prefix)/bin/claude-squad" "$(brew --prefix)/bin/cs"

Manual

Claude Squad can also be installed by running the following command:

curl -fsSL https://raw.githubusercontent.com/smtg-ai/claude-squad/main/install.sh | bash

This puts the cs binary in ~/.local/bin.

To use a custom name for the binary:

curl -fsSL https://raw.githubusercontent.com/smtg-ai/claude-squad/main/install.sh | bash -s -- --name <your-binary-name>

Prerequisites

Usage

Usage:
  cs [flags]
  cs [command]

Available Commands:
  completion  Generate the autocompletion script for the specified shell
  debug       Print debug information like config paths
  help        Help about any command
  reset       Reset all stored instances
  version     Print the version number of claude-squad

Flags:
  -y, --autoyes          [experimental] If enabled, all instances will automatically accept prompts for claude code & aider
  -h, --help             help for claude-squad
  -p, --program string   Program to run in new instances (e.g. 'aider --model ollama_chat/gemma3:1b')

Run the application with:

cs

NOTE: The default program is claude and we recommend using the latest version.


Using Claude Squad with other AI assistants:

  • For Codex: Set your API key with export OPENAI_API_KEY=<your_key>
  • Launch with specific assistants:
    • Codex: cs -p "codex"
    • Aider: cs -p "aider ..."
    • Gemini: cs -p "gemini"
  • Make this the default, by modifying the config file (locate with cs debug)

Menu

The menu at the bottom of the screen shows available commands:

Instance/Session Management
  • n - Create a new session
  • N - Create a new session with a prompt
  • D - Kill (delete) the selected session
  • ↑/j, ↓/k - Navigate between sessions
Actions
  • ↵/o - Attach to the selected session to reprompt
  • ctrl-q - Detach from session
  • s - Commit and push branch to github
  • c - Checkout. Commits changes and pauses the session
  • r - Resume a paused session
  • ? - Show help menu
Navigation
  • tab - Switch between preview tab and diff tab
  • q - Quit the application
  • shift-↓/↑ - scroll in diff view

Configuration

Claude Squad stores its configuration in ~/.claude-squad/config.json. You can find the exact path by running cs debug.

Profiles

Profiles let you define multiple named program configurations and switch between them when creating a new session. When more than one profile is defined, the session creation overlay shows a profile picker that you can navigate with /.

To configure profiles, add a profiles array to your config file and set default_program to the name of the profile to select by default:

{
  "default_program": "claude",
  "profiles": [
    { "name": "claude", "program": "claude" },
    { "name": "codex", "program": "codex" },
    { "name": "aider", "program": "aider --model ollama_chat/gemma3:1b" }
  ]
}

Each profile has two fields:

FieldDescription
nameDisplay name shown in the profile picker
programShell command used to launch the agent for that profile

If no profiles are defined, Claude Squad uses default_program directly as the launch command (the default is claude).

FAQs

Failed to start new session

If you get an error like failed to start new session: timed out waiting for tmux session, update the underlying program (ex. claude) to the latest version.

How It Works

  1. tmux to create isolated terminal sessions for each agent
  2. git worktrees to isolate codebases so each session works on its own branch
  3. A simple TUI interface for easy navigation and management

License

AGPL-3.0

Star History

Star History Chart

Files in the repo

Repository payload24 top-level entries
  • .github
  • app
  • assets
  • cmd
  • config
  • daemon
  • keys
  • log
  • session
  • ui
  • web
  • .gitignore
  • .goreleaser.yaml
  • bump-version.sh
  • CLA.md
  • clean_hard.sh
  • clean.sh
  • CONTRIBUTING.md
  • go.mod
  • go.sum
  • install.sh
  • LICENSE.md
  • main.go
  • README.md

Discussion (0)

Ask about usage, or say what you built with it

Sign in to join the discussion.

No comments yet. Be the first to say what this is good for.

More tools

JuliusBrussee/
caveman

🪨 why use many token when few token do trick — Claude Code skill that cuts 65% of tokens by talking like caveman

105k
1 add
MemPalace/
mempalace

The best-benchmarked open-source AI memory system. And it's free.

59k
stablyai/
orca

Orca is the ADE for working with a fleet of parallel agents. Run any coding agent with your own subscription. Available on desktop, mobile and remote runtime.

66k

A cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, OpenClaw, Grok Build & Hermes Agent. Only official website: ccswitch.io

132k

Never stop coding. Free MIT AI gateway: one endpoint, 352 providers (150+ free), 1200+ models Kimi, Claude, GPT, Gemini, GLM, DeepSeek, MiniMax. Works with Claude Code, Codex, Cursor, OpenCode, Cline & Copilot. Quota-aware auto-fallback, RTK+Caveman compression saves 15-95% tokens, MCP/A2A, Desktop/PWA. Built by 550+ contributors

64k
headroomlabs-ai/
headroom

Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.

71k