This seems 8th kyu easy. Am.... I.... EVOLVING?!?!?

The assertion messages display the inputs clearly, actual and expected results clearly, closing.

This suggestion cannot be implemented under the constraints of different programming language's paradigm. Also, the only thing you can restrict when it comes to lifting up the challenge to the next level is probably code length, which fits well in a code-golf task, but not here, so closing.

Not a kata suggestion. OP solved it, closing...

I made it hard for myself, i filter a[] and b[] for any unique numbers via two sets. (Maybe a new kata?)
Who can read is clearly in the advantage :D
TLDR: Read instructions properly and someone please make the test message less ambigous :)

Both name (arrayDiff / array_diff) is fine for the tests. Not a kata issue.

Javascript: Tests assert for array_diff, boilerplate names function as arrayDiff, should be array_diff

Please use code formatting when posting code. See https://docs.codewars.com/training/troubleshooting#post-discourse.

Some extended test are wrong, please review them

Expected x to equal y indicates that your function returned x when it should have returned y, not the other way around. Not a kata issue.

Some extended test are wrong, please review them

Testing for arrayDiff([0,-12,20,1,20,-15,15,-8,-19,-19,1,2,3,17,-3,-5,19,-16,-18],[0,-12,20,1,20,-15,15,-8,-19,-19,1,2,3,17,-3])
Should work for random arrays too: expected [ +0, -5, 19, -16, -18 ] to deeply equal [ -5, 19, -16, -18 ]

0 it is in both arrays this shouldn't be in the array result, why are you expected 0?

Uhh... Is this really a yellow kata?

The solution is literally a one-liner.

learning this stuffs anyone who like to join me

Olá Luca, o seu código apresenta um problema ao assumir que a variável 'b' possui apenas um elemento. Na verdade, 'b' pode conter múltiplos valores. Consoante você tem no seu código, você acessa apenas o primeiro índice da variável 'b'.