The rules are incomprehensible, even if you read the examples, you need to read until the very end to actually grasp what the kata wants from you.

Why is the expected result for an empty string is an empty array, not an array with a single empty element? There's exactly one way to make an empty string - with zero elements. This choice is so irritating, it's not the first time I see this in a kata that has an empty string as a special case. An empty string usually should NOT be a special case actually. I think returning either [] or [[]] should be allowed for the sake of backwards compatibility, but it should've been only the second one, really. The description speficication of one or more is really weird to me, like why? This doesnt make any sense.

I just tried it and it doesnt work. Why did it not get invalidated?

Fixed. Moved from preloaded to both test suites.

As I recently mentioned on discord, I changed my mind. Do what you think is the best

Bruteforce is not the right approach at all. Try something clever which does not generate a bunch of stuff that is gonna be rejected anyway. Generate only the increasing numbers perhaps, not all of them

non-decreseasing order sounds better, than increasing order with possible duplicates. Maybe you could change that to make the description clearer

Python, probably other languages: no actual random tests. The fixed tests are so pathetic that the code only needs 15 cells of memory. This kata needs to be nuked and redone for all languages that have this kind of tests.

It's no longer there, I suggest we forget that it was. Whatever, I don't thihnk that for a 5 kyu it matters that much