Seems reasonable to me.

(cmd macros didn’t exist when py_str was introduced, IIRC)

r"regex" is used for wrapping "another language". So, it's not clear to me if @*_cmd should always be used for using other languages. I agree that it could be a reasonable convention to have. But it doesn't seem like we have a consistent convention yet. Unless there is a strong convention, I'm not sure if introducing (future) breaking change is a good idea.

Also, @py_cmd could be used for constructing a lazy object that holds a piece of Python code and Julia objects to be injected. This way, run(py`...`) and py"..." would be equivalent. It might be handy for tweaking evaluation contexts (e.g., namespace dict, top-level vs function vs async function).

Thanks for chiming in! I understand your concerns, that it is probably best to establish a standard first, before we make breaking changes. What would you suggest as a path forward for establishing a standard for nested highlighting like this? Is this something that should maybe be a Julep? I do think the ``` for command literals would be a pretty nice convention for this, since it's very reminiscent of how syntax highlighting works in Markdown. I do understand that having ``` be lazy might be useful as well, I don't really have a strong opinion on this.