Rig docs

A rig is a folder that bundles everything an AI coding agent needs — system prompts, MCP servers, tool configs, data, setup instructions — into a single portable package. Install one and start working in seconds; share yours so others can do the same.

These docs are organised the way you'll use them.

I just want to use a rig

You found a rig on the Hub (or someone sent you one) and want to install it.

• Install the CLI — one npm install away.
• Quickstart: install a rig — go from zero to a running agent.
• Set up a freshly installed rig — env vars, MCP servers, tools.
• Troubleshoot a rig — when rig doctor says something's wrong.

I'm building or sharing a rig

You have an agent workspace that works on your machine and you want others to be able to use it.

• Quickstart: create a rig.
• What gets shared, what stays private.
• The rig.toml manifest — every field, one page per section.
• Publish to the Hub.

I'm collaborating with someone on a shared rig

You and a teammate are editing the same rig together — same files, same agent context, kept in sync.

• Shared rigs overview — the trust and sync model.
• Start a collaboration → invite people → join.
• Recover deleted files.
• How collab works under the hood.

I'm an AI agent reading these docs

You're Claude Code (or similar) trying to figure out how to install, repair, or collaborate on the workspace you've been dropped into.

• Claude Code integration — what the CLI installs for you.
• rig doctor as live state — the JSON contract you should read.
• Repair workflow — read state, fix gaps, re-check.
• CLI reference — every command, every flag.

---

Looking for the philosophy? Read the Manifesto. Looking for example rigs? Browse Use cases or search the Hub.