This is a really interesting idea. I wonder if something like Luau would be a good solution here - it's a typed version of Lua meant for sandboxing (built for Roblox scripting) that has a lot of guardrails on it.

https://luau.org/

They use bash in ways a human never would, and it seems very intuitive for them.

Being unchanged for decades means that the training data should provide great results even for the smaller models.

Agents really do not care at all how "nice" a language is. You only need to be picky with language if a human is going to be working with the code. I get the impression that is not the use case here though

I've had LLMs write some pretty complex powershell on the fly. Still a shell language but a lot nicer.

Ideally something like nushell but they don't know that well

just-bash comes with Python installed, so in a way that's what this has done. I've used this for some prototypes with AI tools (via bash-tool), can't really productionise it in our current setup, but it worked very well and was undeniably pretty cool.

Bash is ubiquitous and is not going away any time soon. Nothing is stopping you from doing the same thing with your favorite language.

At least for me codex seems to write way more python than bash for general purpose stuff

I feel like Deno would be perfect for this because it already has a permissions model enforced by the runtime