Well,one cannot make everyone happy. If you are into creating such a sequence good luck doing it. I don't see a reason this is bad. My initial intention was to create a kata for the ones who know about ubbi-dubbi, and provide some fun experience.

I hate to be geeky, but the description says "The rules are simple everytime you see "a" or "e" or "i" or "o" or "u" in a word just prefix it with "ub"." No word about uppercase.

so this i found a small mistake in your test case

function add(a,b) { return a+b; }
var add_ = defaultArguments(add,{b:9});
Test.assertEquals(add_(10), 19);
Test.assertEquals(add_(10,5), 15);

This here
var add_ = defaultArguments(add_,{b:3});

should have not a add_ as an argument but just add because add_ was already created var add_ = defaultArguments(add,{b:9});
that's why I am getting NaN in my final test.

Test.assertEquals(add_(10), 13);

the Same thing is probably in final test case set, I would tell for sure but Kata doesn't let me through;

Yes Tests won't pass at all I don't know maybe it is es6 but my jsbin tells me I did my code right. This was a fair challenge but for tests. So One more time - testing your cases by hand I have absolutely correct results but half an hour ago the last one of final tests told me expected - 13 got 19 and after another half an hour he said he gets NaN. Though I don't get those in my console. Do something with your tests finally!!

pretty lame kata; because you don't actually write a function which is then automatically applied. You pretend as if you really have a given var - questions which is an array. I did everything right, but couldn't pass tests, just because I expected my function to be run on a whatever var automatically - this approach us quite common for the codewars kata. So please review your Kata and make amendments. because the task is ridiculously easy but not with the tests