Skip to content
Back to search

openharness

MarketplaceWorks with Claude CodeGitHub

Harness for agents: run Claude Code, Codex and other coding agents in parallel, wait for them and read their answers, while the person watches every one.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today

Check it deeper on your machine

It has changed since its scan on 15 Nov 2026.

As a Claude Code plugin, install it and then ask Claude Code:

/plugin marketplace add keithmackay/mackayi/plugin install scanskillsafety@mackayiDo a deep scan of https://github.com/autonomous-ai/openharness with scanskillsafety

Or from a terminal, clone it and run it:

git clone https://github.com/keithmackay/scanskillsafety-skill ~/.claude/skills/scanskillsafetynode ~/.claude/skills/scanskillsafety/cli.cjs --deep 'https://github.com/autonomous-ai/openharness'

It checks the exact commit: every file, its packages against OSV (api.osv.dev), how it compares with this page's scan (findsafeskills.com receives the repository's address), and, if you agree, Claude reading every file with your own Claude subscription. It never runs the code.

Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/autonomous-ai--openharness/badge.svg)](/listings/autonomous-ai--openharness)

HTML

<a href="/listings/autonomous-ai--openharness"><img src="/api/listings/autonomous-ai--openharness/badge.svg" alt="findsafeskills safety rating" /></a>

Install

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

.claude-plugin/marketplace.json · Manifest · 24 lines
  1. {
  2. "name": "openharness",
  3. "description": "Harness for agents: run Claude Code, Codex and other coding agents in parallel, wait for them and read their answers, while the person watches every one.",
  4. "owner": {
  5. "name": "Autonomous",
  6. "url": "https://harness.autonomous.ai"
  7. },
  8. "plugins": [
  9. {
  10. "name": "harness",
  11. "source": "./plugins/harness",
  12. "description": "Start, wait for, read and close other coding agents through Harness: the harness skill and Harness's MCP server.",
  13. "category": "development",
  14. "tags": ["agents", "parallel", "codex", "claude-code", "orchestration"]
  15. },
  16. {
  17. "name": "harness-device",
  18. "source": "./plugins/harness-device",
  19. "description": "Shows Claude Code's waiting permission prompts on the Harness round dial, and takes a prompt said at the dial.",
  20. "category": "development",
  21. "tags": ["harness", "dial", "permissions", "voice", "claude-code"]
  22. }
  23. ]
  24. }
README.md · README · 309 lines · 2 findings marked
  1. <p align="center">
  2. <img src="desktop/assets/app_icon.png" width="88" alt="">
  3. </p>
  4. <h1 align="center">Harness like a boss.</h1>
  5. <p align="center">
  6. <b>The ultimate harness for coding agents and beyond.</b><br>
  7. Claude Code at work. Codex at home. Hermes in the cloud. One command center.<br>
  8. Start with code. Then follow your curiosity and build across disciplines: CAD, circuits, robots, games and music.
  9. </p>
  10. <p align="center">
  11. <a href="https://harness.autonomous.ai/desktop?utm_source=github&utm_medium=readme"><b>Download</b></a> ·
  12. <a href="#how-it-works">How it works</a> ·
  13. <a href="#beyond-code">Beyond code</a> ·
  14. <a href="#domain-specific-harnesses-dsh">Harnesses</a> ·
  15. <a href="#harness-device">Device</a>
  16. </p>
  17. ### Every agent, side by side
  18. Claude Code, Codex, Cursor, OpenCode, Devin, Amp, Copilot and seven more.
  19. <p align="center"><img src=".github/assets/readme/agents.gif" width="960" alt="Four agents working at once in one window: Claude Code and Codex on a MacBook, Cursor on an office desktop, OpenCode on a GPU box. Each pane shows its machine, project and branch."></p>
  20. ### Every machine, side by side
  21. Your laptop, home server and GPU box in one window. Link each one with a password.
  22. No SSH keys. No Tailscale. No port forwarding.
  23. <p align="center"><img src=".github/assets/readme/machines.gif" width="960" alt="A new GPU box runs four setup commands and comes online. The app links it with its password, then starts Claude Code there."></p>
  24. ### Keyboard first
  25. ⌘P finds any harness on any machine. ⇧⌘I jumps to the harness waiting on you. ⌘D splits. Every key remaps.
  26. <p align="center"><img src=".github/assets/readme/keyboard.gif" width="960" alt="Keyboard only: open a harness by typing a few letters, zoom it, jump to the agent asking a question and answer it, then split a new pane below."></p>
  27. ### End-to-end encrypted
  28. Code, keys and keystrokes are sealed on your machine. The relay forwards bytes it can't read.
  29. <p align="center"><img src=".github/assets/readme/e2ee.gif" width="960" alt="On the left, an agent rotates a secret and redeploys. On the right, the same harness as the relay sees it: numbered frames of ciphertext."></p>
  30. ### Fast and light
  31. A native app, not Electron. Close it and your agents keep working.
  32. Idle measurements on an M2 Max:
  33. - **⌘N, ⌘O, ⌘T UI:** 11–13 ms median, 15–18 ms p95.
  34. - **Local terminal echo:** 1.1 ms median, 7.2 ms p95, excluding UI rendering.
  35. - **Desktop resources:** ~0.1% of one CPU core and 304 MiB, including terminal rendering and scrollback; agent CLIs and the daemon are excluded.
  36. Desktop results use a Release fixture with 16 terminals and 1,000 scrollback lines each.
  37. See the [workflow and resource benchmarks](docs/performance/2026-09-23-core-experiences.md)
  38. and [remote P2P, TURN and relay results](docs/performance/2026-09-23-transport-routes.md)
  39. for workloads, slow tails, connection failures and raw data.
  40. ### Built the way developers work
  41. - **Real terminals.** Every agent runs in its own tmux pane. Scrollback, colors and keys just work.
  42. - **Your CLIs, as they are.** Harness never wraps an agent. It reads transcripts and uses the vendor's own hooks.
  43. - **A worktree per harness.** Start an agent on its own branch. Your working copy stays clean.
  44. - **Remap every key.** One JSONC file, reloaded on save. Chords up to four strokes.
  45. - **Local models.** Run open-weight models on your own machines with Grid, Ollama, MLX-LM and vLLM.
  46. - **Bytes, not pixels.** Remote terminals stream text peer to peer. No remote desktop.
  47. - **Open source, all of it.** App, CLI, daemon, relay and device, in this repo.
  48. ## How it works
  49. In the workspace, a **harness** is one running session of an agent such as Codex
  50. or Claude Code, with its own conversation and working context. A **tab** groups
  51. harnesses. Use **New Harness** to start one and **New Tab** to group work.
  52. Harnesses in the same tab can message each other with `harness send`: **Agent Messages**,
  53. an experiment of your account, off by default and turned on in Settings ▸ Experimental.
  54. The Store offers **harnesses** with instructions, tools, and optional viewers for
  55. specific crafts. Install a harness, then start it in your workspace. See the
  56. [Naming System](docs/naming-system.md) for the complete naming rules.
  57. One daemon per machine runs your agents in tmux. It dials out, so no machine opens a port.
  58. ```mermaid
  59. flowchart LR
  60. subgraph laptop["Your laptop"]
  61. app["Harness app"] -- loopback --> d1["daemon"]
  62. d1 --> a1["tmux · Claude Code · Codex"]
  63. end
  64. subgraph box["Your GPU box"]
  65. d2["daemon"] --> a2["tmux · OpenCode · Hermes"]
  66. end
  67. device["Harness device"] -- USB --> d1
  68. d1 <-. "direct WebRTC" .-> d2
  69. d1 <-- "ciphertext" --> relay["Harness relay · no keys"]
  70. relay <-- "ciphertext" --> d2
  71. ```
  72. Every path is sealed end to end: ChaCha20-Poly1305, X25519 session keys, pinned Ed25519 identities.
  73. Harness picks the best path on its own.
  74. **Direct.** A WebRTC channel between your machines. No server in the path.
  75. <p align="center"><img src=".github/assets/readme/connect/direct.gif" width="800" alt="The Harness window connects straight to home-server, office-desktop, cloud-server and mac-mini over WebRTC, with no server in the path."></p>
  76. **Through Cloudflare.** When a firewall blocks the direct path, the same encrypted channel runs over
  77. Cloudflare's TURN network. Harness keeps trying for a direct path.
  78. <p align="center"><img src=".github/assets/readme/connect/cloudflare.gif" width="800" alt="The encrypted WebRTC channel passes through the nearest Cloudflare edge when a firewall blocks the direct path."></p>
  79. **Through our relay.** A fallback while WebRTC negotiates. The relay holds no keys and forwards ciphertext.
  80. <p align="center"><img src=".github/assets/readme/connect/relay.gif" width="800" alt="The Harness relay keeps the harness reachable over its WebSocket while WebRTC negotiates. It holds no keys."></p>
  81. The [architecture guide](docs/architecture.md) has the details.
  82. <a id="run-it"></a>
  83. ## Get started
  84. **[Download the app](https://harness.autonomous.ai/desktop?utm_source=github&utm_medium=readme)** for macOS or Linux.
  85. Add a machine. Run this on it, then **Machines → Link Machine** in the app:
  86. ```bash
  87. curl -fsSL https://harness.autonomous.ai/cli/install.sh | bash
    • Note: Pipes a downloaded script directly into a shell (curl | sh / wget | bash)
    • Line 127 on GitHub
  88. harness login
  89. harness remote-password set
  90. harness start
  91. ```
  92. **In a terminal:** the same line installs `hn` — tmux's keys and your `~/.tmux.conf`, with every
  93. harness on every machine ([hn](tui/README.md)). The first time, `hn` signs in and connects the
  94. computer:
  95. ```bash
  96. curl -fsSL https://harness.autonomous.ai/cli/install.sh | bash
    • Note: Pipes a downloaded script directly into a shell (curl | sh / wget | bash)
    • Line 138 on GitHub
  97. hn
  98. ```
  99. **For agents:** Claude Code and Codex can start, wait for and read other harnesses with the `harness` skill
  100. ([harness.autonomous.ai/agents](https://harness.autonomous.ai/agents), which has the Codex lines too):
  101. ```bash
  102. claude plugin marketplace add autonomous-ai/openharness --sparse .claude-plugin plugins
  103. claude plugin install harness@openharness
  104. ```
  105. <details>
  106. <summary><b>Build from source</b></summary>
  107. Needs Node.js 20+, tmux, Xcode and Flutter 3.47+ / Dart 3.13+:
  108. ```bash
  109. git clone https://github.com/autonomous-ai/openharness.git
  110. cd openharness
  111. (cd cli && npm ci)
  112. make install-cli
  113. cd desktop
  114. flutter config --enable-swift-package-manager
  115. flutter pub get
  116. flutter run -d macos
  117. ```
  118. `make install-cli` installs this checkout's CLI and restarts the local daemon. See the
  119. [development guide](docs/development.md).
  120. </details>
  121. <a id="beyond-code"></a>
  122. ## Beyond code: Build across disciplines
  123. > “World-class entrepreneurs are polymaths.” — [Peter Thiel](https://www.youtube.com/watch?v=h10kXgTdhNU&t=811s)
  124. Coding agents can build far more than software. Give one a harness and it works with the real
  125. tools of a craft. You steer in a live viewer. Every clip below is a real session.
  126. ### Beyond code: Design
  127. **[Blender](store/agents/blender/).** Ask for a lamp and the sliders that matter. Turn them and Blender rebuilds the geometry. Keep the versions you love.
  128. <p align="center"><img src=".github/assets/readme/beyond/blender.gif" width="800" alt="Shape Lab in Blender: dragging height and twist sliders rebuilds a ribbon lamp, and chosen designs are kept."></p>
  129. ### Beyond code: Circuits
  130. **[CircuitJS](store/agents/circuitjs/).** Build a filter. Change one resistor. Overlay the new trace, measure the difference and keep both.
  131. <p align="center"><img src=".github/assets/readme/beyond/circuitjs.gif" width="800" alt="Scope Lab in CircuitJS: an RC filter captured at 1 kΩ and 2 kΩ, traces overlaid and measured with cursors."></p>
  132. ### Beyond code: Robotics
  133. **[MuJoCo](store/agents/mujoco/).** Pin a moment in a robot's run. Shove it with 100 N. Watch two futures split.
  134. <p align="center"><img src=".github/assets/readme/beyond/mujoco.gif" width="800" alt="A Unitree Go2 in MuJoCo: the original and shoved futures play together with a height chart."></p>
  135. ### Beyond code: Games
  136. **[Godogen](store/agents/godogen/).** Your agent makes a playable game. Miss a jump, rewind, try again. Pin the moment so the agent sees what you mean.
  137. <p align="center"><img src=".github/assets/readme/beyond/godogen.gif" width="800" alt="Alpine Drift, a game made with Godogen: a run is rewound, retried and a moment is pinned with feedback."></p>
  138. ### Beyond code: Music
  139. **[Strudel](store/agents/strudel/).** The track is code you can perform. Bring voices in and out, mark the good parts, keep the WAV. [Hear it](https://github.com/user-attachments/assets/a3d4381b-5f55-406c-9d68-330cd8792fc5).
  140. <p align="center"><img src=".github/assets/readme/beyond/strudel.gif" width="800" alt="A live Strudel performance: voice lanes play beside the code, and the take is kept with markers."></p>
  141. ### Beyond code: Chemistry
  142. **[RDKit](store/agents/rdkit/).** Turn a bond and watch the molecule move. Follow the real energy curve. Keep the pose worth a closer look.
  143. <p align="center"><img src=".github/assets/readme/beyond/rdkit.gif" width="800" alt="A bond scan in RDKit: the molecule rotates through sampled poses along an MMFF94 energy curve."></p>
  144. ### Beyond code: Documents
  145. **[Typst](store/agents/typst/).** Your agent writes a real PDF. Circle a detail, quote a line, leave a note. The next draft answers it.
  146. <p align="center"><img src=".github/assets/readme/beyond/typst.gif" width="800" alt="A Typst PDF under review: notes are pinned to an area and a sentence, then carried to the next draft."></p>
  147. ## Domain-specific harnesses (DSH)
  148. Code is the common medium. Geometry scripts make parts. Netlists make boards. Animation code makes film.
  149. A **domain-specific harness** turns a coding agent into a specialist. It brings the craft's
  150. instructions and skills, a pinned toolchain, a project template, checks and a **live viewer**.
  151. You chat on one side. The board, the part or the game takes shape on the other.
  152. The agent does the reasoning. The harness supplies the tools and the view. It's a folder with a
  153. `harness.json`, so adding a craft never touches the app.
  154. <!-- store-catalog:start -->
  155. ### 49 harnesses in the Store
  156. | Category | Agents and harnesses |
  157. |---|---|
  158. | **Coding** | [Claude Code, Codex, Cursor, OpenCode, Pi, Hermes, Command Code, Devin, Muse Code, Amp, Antigravity, GitHub Copilot, Grok Build, Kilo Code](docs/engines.md), [Machine Monitor](store/agents/machine-monitor/), [Trail of Bits Skills](store/agents/trailofbits-skills/) |
  159. | Essentials | [Harness Monitor](store/agents/harness-monitor/), [Memories](store/agents/memories/) |
  160. | Design | [Autonomous Workshop](store/agents/autonomous-workshop/), [Blender](store/agents/blender/), [Bonsai MCP](store/agents/bonsai-mcp/), [Creative Direction](store/agents/creative-direction/), [Excalidraw](store/agents/excalidraw/), [FreeCAD](store/agents/freecad/), [Generative Art](store/agents/generative-art/), [OpenSCAD](store/agents/openscad/), [text-to-cad](store/agents/text-to-cad/) |
  161. | Engineering | [Autonomous Circuit](store/agents/autonomous-circuit/), [CircuitJS](store/agents/circuitjs/), [Home Assistant](store/agents/home-assistant/), [KiCad](store/agents/kicad/), [Orca Slicer](store/agents/orca-slicer/), [Yosys](store/agents/yosys/) |
  162. | Media | [Comfy MCP](store/agents/comfy-mcp/), [Manim](store/agents/manim/), [OpenMontage](store/agents/openmontage/), [Remotion](store/agents/remotion/) |
  163. | Music | [Ableton AI](store/agents/ableton-ai/), [JUCE Agent Toolkit](store/agents/juce-agent-toolkit/), [Music Studio](store/agents/music-studio/), [Score](store/agents/score/), [Strudel](store/agents/strudel/) |
  164. | Productivity | [Marp](store/agents/marp/), [Typst](store/agents/typst/) |
  165. | Science & Data | [autoresearch-mlx](store/agents/autoresearch-mlx/), [Data Studio](store/agents/data-studio/), [Lab Bench](store/agents/lab-bench/), [marimo](store/agents/marimo/), [RDKit](store/agents/rdkit/) |
  166. | Simulation | [DimOS](store/agents/dimos/), [Drone Pilot](store/agents/drone-pilot/), [Foam-Agent](store/agents/foam-agent/), [MuJoCo](store/agents/mujoco/), [SimSkill](store/agents/simskill/) |
  167. | Games | [Game Master](store/agents/game-master/), [Godogen](store/agents/godogen/), [Phaser](store/agents/phaser/), [Voxel Worlds](store/agents/voxel-worlds/) |
  168. | Research | [Roundtable](store/agents/roundtable/) |
  169. | Local AI | [MLX-LM](store/agents/mlx-lm/), [Model Manager](store/agents/autonomous-grid/), [Ollama](store/agents/ollama/), [vLLM](store/agents/vllm/) |
  170. Upstream open-source tools and original workflows. 10 [shared viewers](store/viewers/) install alongside
  171. the harnesses that need them. Unlisted experiments are not shown.
  172. <!-- store-catalog:end -->
  173. ### Build your own
  174. The next harness is the one for your craft. Wrap a tool you love or your team's toolchain.
  175. It can live here or in your own repo. From this repo:
  176. ```json
  177. {
  178. "spec": 1,
  179. "id": "examples/hello-world",
  180. "name": "Hello World",
  181. "engine": "codex",
  182. "workspace": { "template": "template", "marker": "index.html" },
  183. "agent": { "instructions": "AGENTS.md" },
  184. "viewer": { "use": "autonomous/web-viewer" }
  185. }
  186. ```
  187. ```bash
  188. harness dsh install "$PWD/store/viewers/web-viewer" --link
  189. cp -R store/examples/hello-world ../my-harness
  190. harness dsh check ../my-harness
  191. harness dsh install ../my-harness --link
  192. ```
  193. Press **⌘N → Hello World** and say hello. The [authoring guide](store/README.md) covers the rest.
  194. ## Harness device
  195. [**Get a Harness device**](https://www.autonomous.ai/harness-device), or build your own from the files below. The
  196. [firmware guide](devices/harness-device/firmware/README.md) lists the supported boards and build
  197. commands, and the [hardware guide](devices/harness-device/hardware/README.md) covers the design files.
  198. <p align="center"><img src=".github/assets/readme/device.gif" width="960" alt="A finger taps the round Harness device, speaks a task to fix the login flow, and the device shows the agent deploying, then a summary of the shipped fix."></p>
  199. The optional **Harness device** is a round, always-on display that sits beside
  200. your keyboard and shows your agents at a glance: what each one is doing, which one has finished, and
  201. which one is waiting on you. Read a question and answer it on the screen, or tap and speak a new task,
  202. without switching windows.
  203. It is open hardware, all the way down. This repository has everything it takes to build one:
  204. | Layer | What's here |
  205. |---|---|
  206. | [Firmware](devices/harness-device/firmware/) | ESP32-S3, ESP-IDF, a 466 × 466 round AMOLED with touch, microphones and audio. Connects to the host's daemon over USB — no Wi-Fi setup, no account on the device. |
  207. | [PCB](devices/harness-device/hardware/pcb/) | The EasyEDA Pro project, the schematic, Gerbers, the bill of materials, and pick-and-place data for assembly. |
  208. | [Enclosure](devices/harness-device/hardware/3d/) | STEP for editing and STL for printing: the housing, an iron counterweight base, the USB clamp, and the button. |
  209. ## Contribute
  210. Make a harness for a tool you love. Improve terminals, engines, the daemon or the relay. Port the
  211. firmware. Start with the [contribution guide](CONTRIBUTING.md).
  212. [Architecture](docs/architecture.md) · [Product direction (proposal)](docs/product-direction.md) ·
  213. [Development](docs/development.md) · [Extending](docs/extending.md) · [CLI](docs/cli.md) ·
  214. [Security](SECURITY.md) · [MIT license](LICENSE); upstream tools keep their own.

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.