Guide · Agent skill

How to use the OpenRig team setup agent skill (rigs)

Updated 2026-11-02 · written with AI from the project's own documentation and checked against our catalogue data

The rigs skill is an instruction package that teaches an AI agent, such as Claude Code or Codex, how to install OpenRig and set up a team of coding agents on your project. This guide explains what the skill does, how to install it, how to ask for it in plain language, and what it changes on your machine. It is written for people who have not used agent skills before.

OpenRig team setupmvschwarz/openrig · Apache-2.0 Download the skill

Quick start

1Download and unzip. Download the rigs ZIP from the directory page and unzip it into ~/.claude/skills/ for all projects, or into .claude/skills/ in one project. In the Claude apps, upload the ZIP under Settings, then Capabilities, then Skills.
2Check your requirements. Confirm Node.js 22 or 24, tmux, and a signed-in Claude Code or Codex are present, using node --version, tmux -V and the login status commands.
3Ask for the setup. Tell your agent you want OpenRig and a team of coding agents on your project, and name the repository. The skill will explain what it will install and change before it runs anything.
4Install OpenRig. Approve the install. The skill runs npm install -g @openrig/cli, then rig setup with your chosen tool, then rig doctor.
5Start the daemon. The skill checks rig daemon status and runs rig daemon start if needed. This also starts OpenRig's kernel team.
6Open the operator and state your goal. The skill opens the OpenRig view with rig terminal open saved:kernel --window. Describe the work you want done, and the operator sets up the team.

What the rigs skill makes the agent do

OpenRig runs a team of coding agents, such as Claude Code and Codex, side by side in tmux. A local background daemon keeps their tasks, messages and state, so the team keeps working across restarts. The rigs skill guides your agent through getting OpenRig installed and running, then hands your goal to OpenRig's operator so it can set up a team on your project.

The skill also covers finding your way back to OpenRig later: seeing your OpenRig agents, re-attaching to them, bringing a stopped team back, and joining your current session to a team. Operating a running team from inside it (recovery, handovers, upgrades) is a separate skill called openrig-skills.

Install the rigs skill

Download the rigs skill ZIP from the directory page and unzip it. You can put the folder in one of two places. For all your projects, use ~/.claude/skills/. For one project only, use .claude/skills/ inside that project's folder.

In the Claude apps you can instead upload the ZIP under Settings, then Capabilities, then Skills.

Personal install: unzip into ~/.claude/skills/
Project install: unzip into .claude/skills/ in your project
Claude apps: Settings → Capabilities → Skills, then upload the ZIP

Ask for it with a natural request

You do not need a special command. Mention OpenRig, rigs or the rig command, or say what you want, such as a team of coding agents on your project. The skill also applies when you ask to install, update, start or re-attach to OpenRig, or to share or join a team.

Say what you want built in plain words. The skill is designed to give that goal to OpenRig's operator rather than building it in your current session.

"Set up OpenRig and a Claude Code and Codex team on this repository."
"Get back to my OpenRig agents."
"Bring my stopped team back."

Know what the setup changes on your machine

Setup is not a sandbox. Starting OpenRig and a team writes files in your home folder and in the repository. These include Codex settings in ~/.codex/config.toml, Claude Code workspace trust in ~/.claude.json, skills in ~/.claude/skills and ~/.agents/skills, and ~/.tmux.conf. In the project, it can write AGENTS.md or CLAUDE.md, and .claude/settings.local.json.

By default, OpenRig's own seats run Codex with -s danger-full-access -a never, and Claude Code with file tools, web fetches and operational commands allowed without asking. The skill tells you this before it starts OpenRig. There is no project-only mode, so if you want changes kept inside one project, read the full list first. The skill also asks for your approval before installing anything or starting a team.

Setup writes shared home-folder files that other Codex or Claude Code sessions can read
Codex seat startup can add workspace trust and config fragments even if setup was skipped
Setup needs network access and runs scripts

Quick check before you start

The skill checks what is missing before it installs anything. You can run the same checks yourself. Setup needs macOS or Linux, Node.js 22 or 24, tmux, and Claude Code or Codex signed in. npm is needed to install the OpenRig command-line package. Homebrew is optional and is used to install tmux when present. Herdr is optional and is installed for the OpenRig view unless you decline it.

node --version
tmux -V
claude auth status
codex login status

Install OpenRig and check it

The skill installs the OpenRig command-line package, then runs setup for the tool you use. Use claude, codex, or both. Setup checks Node, tmux and your logins, and installs tmux or the chosen tool if missing. It adds an OpenRig block to ~/.tmux.conf and does not start the daemon. Then it runs a doctor check.

If rig will not run, the help guide for the installed version is available with rig context get help. Getting-started notes are at openrig.dev/docs/getting-started.

npm install -g @openrig/cli
rig setup --providers codex
rig doctor

Start OpenRig and open the operator

Once OpenRig is installed, the skill checks whether the daemon is running and starts it if needed. Starting the daemon also starts OpenRig's own team, called the kernel, whose operator sets up your team. Then the skill opens the OpenRig view for you, so you can see the operator and give it your goal.

The same command is also how you get back to your agents later. The skill explains each change before running it.

Use rig terminal open saved:kernel --window to open the view again later
Under 120 columns the operator fills the first page; wider windows show the dashboard and operator together

rig daemon status
rig daemon start
rig terminal open saved:kernel --window

What people use it for

Set up a Claude Code and Codex team that coordinates on one repository
Bring back a team that was stopped, without rebuilding it
Re-attach to your OpenRig agents from a new terminal session
Keep a multi-agent team running across machine restarts through the local daemon

Frequently asked questions

Is the rigs skill free to use?

The skill is published under the Apache-2.0 licence, as is the OpenRig project it installs. You still need your own Claude Code or Codex access, since OpenRig runs those tools.

What do I need before installing the rigs skill?

You need macOS or Linux, Node.js 22 or 24, tmux, npm, and Claude Code or Codex signed in. Homebrew and Herdr are optional. The skill checks what is missing before it installs anything.

Where do I put the rigs skill folder?

Unzip it into ~/.claude/skills/ to use it across projects, or into .claude/skills/ inside one project. In the Claude apps, upload the ZIP under Settings, then Capabilities, then Skills.

Will this change my Codex or Claude Code settings?

Yes. Setup writes to shared files such as ~/.codex/config.toml, ~/.claude.json and ~/.tmux.conf, and there is no project-only mode. If you want changes kept inside one project, read the full list of changes before you approve.

How is OpenRig different from running one agent for a small task?

OpenRig installs and runs a whole multi-agent environment with a background daemon that keeps tasks and messages across restarts. For a small single-agent task, it is more than you need.

What if the install fails or rig will not run?

Run rig context get help to read the help guide for your installed version. If rig itself will not start, see openrig.dev/help/agents, and the getting-started notes at openrig.dev/docs/getting-started.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

Use the rigs skill if you already use Claude Code or Codex and want several agents coordinating on one project that survives restarts and can be re-attached later. Pick something else if you want only one agent for a small task, or if you do not want OpenRig writing to your shared Codex, Claude Code and tmux configuration.

Download the skill