Ah, so there's no need to have them as local variables of the test group functions?

I did some changes to test cases. The names of the randomized name functions in test cases I didn't modify, those were like that to begin with.

Alright. I will need to take a break to do something for a while, but will go ahead after that.

I've changed the example test cases. I would like to know whether this seems right or whether I should fix anything, and whether to make similar changes to the actual tests as well.

Yes, I did use "fork". I'll take a look at the link and see what I'd need to do.

https://www.codewars.com/kumite/61e3e275af28c2005c57c42d
I believe this fixes the "NameError: name 'Test' is not defined" error for the sample tests with Python 3.8.

That is fair. Do I just avoid solving the easier Kata then, or something else?

https://www.codewars.com/kumite/61e083bde8c73e00579ca983
I hope this would prevent random tests with n = 1 from appearing (for Python 3.8), which the Kata description seems to suggest shouldn't occur.