A CLI and workflow engine that runs coding agents like Claude Code, Codex and Copilot through configurable review, repair and test gates before work is accepted.
mvschwarz/openrig
A TypeScript CLI and terminal UI for defining, launching and running persistent teams of Claude Code and Codex agents in tmux.
- スター
- 7,843
- フォーク
- 624
- 未解決の Issue
- 167
- 最終コミット
- 2 時間前
これは何?
OpenRig is an open-source harness for coordinating multiple AI coding agents. You describe a team of roles and pods in a YAML RigSpec, then boot it with a single command. Each agent runs in its own tmux session, and the team can mix Claude Code and Codex in the same rig. The project provides a local daemon, a CLI, a terminal UI and an MCP server so agents and people can inspect state, send messages and track owned work.
The main problem it addresses is terminal sprawl: many independent agent sessions with no shared structure. OpenRig keeps stable seat addresses such as dev-build@starter, so a role persists even when the conversation in it changes. It can snapshot a topology on shutdown and restore it by name, discover and adopt existing tmux sessions, and route messages between seats with rig send and rig broadcast.
OpenRig requires Node.js 22 or 24 and tmux, and runs on macOS or Linux (Windows through WSL2). Launching a rig writes provider hook and workspace trust configuration, so the documentation's setup and permission notes should be read before first use.
主な機能
- 01Defines multi-agent teams declaratively in YAML with pods, edges and continuity policies.
- 02Runs Claude Code and Codex agents side by side, each in its own tmux session.
- 03Snapshots a topology on shutdown and restores it by name.
- 04Discovers existing tmux agent sessions and adopts them into a managed rig.
- 05Provides rig send and broadcast messaging between seats, plus a terminal dashboard and an MCP server.
ユースケース
- Coordinating a builder and a reviewer agent working on one repository.
- Keeping a long-running set of coding agents organized and recoverable after a reboot.
- Standing up a reusable team template that can be shared as a bundle.
トピック
よくある質問
- openrig とは?
- A TypeScript CLI and terminal UI for defining, launching and running persistent teams of Claude Code and Codex agents in tmux. OpenRig is an open-source harness for coordinating multiple AI coding agents. You describe a team of roles and pods in a YAML RigSpec, then boot it with a single command. Each agent runs in its own tmux session, and the team can mix Claude Code and Codex in the same rig. The project provides a local daemon, a CLI, a terminal UI and an MCP server so agents and people can inspect state, send messages and track owned work.
- openrig は無料で使えますか?ライセンスは?
- openrig は Apache-2.0 ライセンスのオープンソースです。商用製品で使う前にライセンス条件を確認してください。
- openrig は今もメンテナンスされていますか?
- 活発にメンテナンス中。openrig の最終コミットは 2026年10月19日 です。
- openrig は何に使えますか?
- Coordinating a builder and a reviewer agent working on one repository. Keeping a long-running set of coding agents organized and recoverable after a reboot. Standing up a reusable team template that can be shared as a bundle.
- openrig はどこで入手できますか?
- ソースコードとインストール手順は GitHub にあります:https://github.com/mvschwarz/openrig
概要はプロジェクトの README をもとに AI が生成しています。GitHub の数値は毎日更新されます。
関連プロジェクト
Multi-model AI coding orchestrator where a coordinator routes Markdown-defined specialist tasks to Claude, Codex, Gemini, Kimi and Grok CLIs in isolated git worktrees.
A mobile, web and desktop app that runs Claude Code, Codex and Grok sessions from one place, using your existing subscriptions.
A VS Code extension that puts Claude Code, Codex and other coding agents in one conversation, so they can give separate opinions, hand off work and debate solutions.
A coding-agent skill that makes AI agents prove their code with a test-plan-first workflow, a verification gauntlet, and an evidence report.
AIエージェントがネイティブバイナリ、Electronアプリ、.NETアセンブリ、Webサイトを解析するMCPサーバー兼CLI。