The other test cases are wrong, I don't know C# but I'll try and fix it when I have time, if the person who translated it doesn't fix it before then

Yes, if you make a translation I will approve it

Ok didn't know about that, I'll fix it now
Fixed.

ah sorry about that facepalm
I'll go fix that now
EDIT: nvm, smile67 just did

I wasn't paying attention when making the random tests, so I accidentally made it generate jagged arrays instead of square arrays.
Are there any problems other than the jagged arrays?

I made a JS translation

Yeah, I thought it might be the case that it wouldn't translate well into JS. JS Proxies look interesting though, hadn't heard of them before

Now that I think of it, it seems more logical to return [1,1], because 1 is also a loop, it's just the loop that defines a happy number. Thoughts?

I agree, I've managed to solve a single 2 kyu kata, but I still often have trouble with 5 kyu katas. And completing the too-easy-for-their-rank ones means I'm at rank 3 kyu now, and I don't really feel like I've earned it. 2 kyu and 1 kyu will definitely be more challenging though, since there are so few available at those ranks.

Needs random tests in JS.

JavaScript version needs random tests.

JavaScript needs random test cases.

This looks interesting, but unfortunately it's only in Ruby. Translating it myself wouldn't be as fun as actually doing it, so if someone who has completed it could translate it to JS that would be great.
(Although I'm not sure how well it will translate, considering Ruby syntax looks quite different than JS)

I made a translation