A strict mandate would not be effective and would possibly be harmful for the reasons pointed out by @ByteCommander in this comment.
My gut reaction is that I could certainly get behind the idea of automating notifying users with a short explanation, with a link to “How do I ask a good question?” as I've lost count of how many times I've encouraged a user to do just that, however in order to make that determination I'd have to actually read the question and at least make an attempt to understand it.
This is a task that's exceedingly hard to do programmatically as humans are still better as sifting through this stuff than code is. There are a number of related questions on how to handle these non-questions on meta.stackexchange.com and after sifting through them I think the most related example is Should non-questions be rewritten as questions? I understand that it's more fun to answer questions than to play janitor and attempt to improve posts like you describe, but it's a necessary task.
In situations such as you mention here, what I try to do is at least point them toward https://askubuntu.com/help/how-to-ask and if I think I get the gist of it, take my best shot at converting it to a question and drop a comment suggesting that the OP revise it if I've misinterpreted them.
While on the surface it appears that using the absence of a single question mark as a trigger could be useful, I think the number of "false positives" could easily outweigh the usefulness of such an approach. For instance the statement "I can't set my script as executable" has no question mark, and clearly isn't a good question but it's pretty clear what the problem is. I don't think we'd have a problem pointing them in the right direction.