Happy to see you getting some non-engineer recognition, simonw!

I feel like somebody is watching my Claude sessions :)

—-

2. we want to use docker for this not qemu if possible.. we should still be able to connect and debug it the same i assume 3. why does it need to be QEMU and not docker for this to be a real exploit? 4. no just use QEMU then

Is this not just the sandbox that Codex runs in? Are you invoking it from the ChatGPT interface, similar to Claude Artifacts?

Even a gift link requires a login. Guess I’m not the target audience.

I’m interested to understand how this is different than Jelly; they seem to be similar. Same for Scriptable. I’ve been looking at this to hand over to Claude to build Shortcuts, something which has a terrible development experience.

You can definitely have Claude work with cherri files.

Jelly was a confusing experience for me, with JellyCuts becoming closed source and focusing on advertising, then Open-Jellycore branching out but not actually keeping up with the latest shortcut actions.

Cherri has almost every action you can find in the Shortcuts app, easy to use, and easy to create Shortcuts that can accept input and output so that they can be automated or scripted further.

You’ll have challenges with this too but you can get something by working with the three top labs’ models. Tried on Arena.ai and sent any errors back (in a personal effort to further iOS accessibility, but I digress).

Wonderful project, thank you Cherri!

Main difference is language style.

Cherri feels like your writing Go.

Jelly feels like your writing JavaScript / Swift.

Using GitHub MFA via the app on my iPhone.

yea. I'm pretty sure they want separate authenticator app or browser extension

So now I need my damn phone to push something. Great. What's next, my national ID?

If by need you mean, can choose to use, and if by push you mean, login to the GitHub web ui, then sure.

lmao welp. that is the path other apps are going so i wouldnt be surprised

Good recent video on this specific subject, it was enlightening to me: https://youtu.be/3FbqaD1MCUA

The product being evaluated is a home security camera agent, its user base is HomeAssistant-adjacent. Value here is privacy over latency (23tok/sec isn’t amazing for a vision model)

I’m on a similar trajectory. Claude is excellent at writing Dockerfiles, compose files, and operating docker. “Add a daily backup container to this compose file which writes to a new volume”, and it’s done.

I also experimented with skill-creator to generate a skill for “operating” my chezmoi in a read only way, and now when I use Claude in a certain directory it tells me what needs to be updated and which commands to issue. I can see why people are worried about human skill decline!

Risk, imo. I am running it, and constantly identify use cases but they each carry some kind of risk I can’t get past. And so for now I am using individual purpose-built Claude Code harnesses, which has much more cognitive overhead and isn’t wired into discord, but I can really control access.

They are centered around managing information in/out, or “hey siri have groceries delivered“.

Yes, I use it that way too. Usually for things like code changes on my project when I'm away from home. Besides that, it does market research for me, but those were the only things I've been able to do so far. I can't find a use case that really stands out.