My current fave harness. I've been using it to great effect, since it is self-extensible, and added support for it to <a href="https://github.com/rcarmo/vibes" rel="nofollow">https://github.com/rcarmo/vibes</a> because it is so much faster than ACP.
Pi has made all the right design choices. Shout out to Mario (and Armin the OG stan) — great taste shows itself.
I’m working with a friend to build an ui around Pi to make it more user friendly for people who prefer to work with a gui (ala conductor). You can check out the repo: <a href="https://github.com/philipp-spiess/modern" rel="nofollow">https://github.com/philipp-spiess/modern</a>
Hugging Face now provides instructions for using local models in Pi:<p><a href="https://x.com/victormustar/status/2026380984866710002" rel="nofollow">https://x.com/victormustar/status/2026380984866710002</a>
Preconfigured PI: <a href="https://github.com/can1357/oh-my-pi" rel="nofollow">https://github.com/can1357/oh-my-pi</a>
<i>Pi ships with powerful defaults but skips features like sub-agents and plan mode</i><p>Does anyone have an idea as to why this would be a feature? don't you want to have a discussion with your agent to iron out the details before moving onto the implementation (build) phase?<p>In any case, looks cool :)<p>EDIT 1: Formatting
EDIT 2: Thanks everyone for your input. I was not aware of the extensibility model that pi had in mind or that you can also iterate your plan on a PLAN.md file. Very interesting approach. I'll have a look and give it a go.
See my comment in the thread but there is an intuitive extension architecture that makes integrating these type of things feel native.<p><a href="https://github.com/badlogic/pi-mono/tree/main/packages/coding-agent#extensions" rel="nofollow">https://github.com/badlogic/pi-mono/tree/main/packages/codin...</a>
I plan all the time. I just tell Pi to create a Plan.md file, and we iterate on it until we are ready to implement.
Check <a href="https://pi.dev/packages" rel="nofollow">https://pi.dev/packages</a><p>There are already multiple implementations of everything.<p>With a powerful and extensible core, you don't need everything prepackaged.
I’ve been testing it for a few days on pretty much clean install (no customizations/extensions) and it’s ok. Not sure if I like it yet.
The way you’re able to extend the harness through extension/hook architecture is really cool.<p>Eg some form of comprehensive planning/spec workflow is best modeled as an extension vs natively built in. And the extension still ends up feeling “native” in use
I've been using Pi day to day recently for simple, smaller tasks. It's a great harness for use with smaller parameter size models given the system prompt is quite a bit shorter vs Claude or Codex (and it uses a nice small set of tools by default).
Has anyone used an open coding agent in headless mode? I have a system cobbled together with exceptions going to a centralized system where I can then have each one pulled out and `claude -p`'d but I'd rather just integrate an open coding agent into the loop because it's less janky and then I'll have it try to fix the problem and propose a PR for me to review. If anyone else has used pi.dev or opencode or aider in this mode (completely non-interactive until the PR) I'd be curious to hear.<p>EDIT: Thank you to both responders. I'll just try the two options out then.
Been using pi exactly for this and it's working great!
fast-agent lets you do this as well (and has a skill in its default skills repo to help with automation/running in container/hf job).
You probably want to look into pi then - it's extremely extensible.
you can run <a href="https://block.github.io/goose/" rel="nofollow">https://block.github.io/goose/</a> in headless mode (I work on goose)
Note there is a fork oh-my-pi: <a href="https://github.com/can1357/oh-my-pi" rel="nofollow">https://github.com/can1357/oh-my-pi</a>
of <a href="https://blog.can.ac/2026/02/12/the-harness-problem/" rel="nofollow">https://blog.can.ac/2026/02/12/the-harness-problem/</a> fame. I use it as a daily driver but I also love pi.
But I can't use my Codex plan with it, right? I have to use an API key?
You can use your Codex plan with it. OpenAI endorsed it several weeks ago, as far as I remember. That could change, however, but now seems safe.
Is that an official term "coding harness"<p>Wondering if you wanted a similar interface (though a GUI not just CLI) where it's not for coding what would you call that?<p>Same idea cycle through models, ask question, drag-drop images, etc...
Yes. It seems to be the term that stands out the most, as terms like "AI coding assistant", "agentic coding framework", etc. are too vague to really differentiate these tools.<p>"harness" fits pretty nicely IMO. It can be used as a single word, and it's not too semantically overloaded to be useful in this context.
LLM harness has been in vogue for a year now…
Anyone managed to run pi in a completely sandboxed environment? It can only access the cwd and subdirectories
I run mine inside <a href="https://github.com/rcarmo/agentbox" rel="nofollow">https://github.com/rcarmo/agentbox</a> (with <a href="https://github.com/rcarmo/webterm" rel="nofollow">https://github.com/rcarmo/webterm</a>)
I do this with an extension. I run all bash tools with bwrap and ACLs for the write and edit tools. Serves my purposes. Opens up access to other required directories, at least for git and rust.<p>I think I published it. Check the pi package page.
I’ve been tinkering with Gondolin, a micro-vm agent sandbox.<p>Here’s an example config: <a href="https://github.com/earendil-works/gondolin/blob/main/host/examples/pi-gondolin.ts" rel="nofollow">https://github.com/earendil-works/gondolin/blob/main/host/ex...</a>
[dead]
[dead]
The backing to OpenClaw/MoltBot whatever they're calling themselves. Why is it insecure, well, Pi tells you >No permission popups.<p>Anyway, even if you give your agent permission, there's no secure way to know whether what they're asking to is what they'll actually do, etc.
Just how expensive was that domain?