That's not an issue I think. Many katas have 'bonus point' followups that are now actually required.

Added in this fork

Added in this fork

resolved for all languages

resolved for all languages (+ resolved the problem with pseudo-random tests as well)

Should probably remove the no argument case across all languages. (It doesn't make sense to find the maximum of an empty list anyway)

Then the description can be changed to something like:

Your job is to complete the function that receives one or more keys as arguments, and returns the longest keyword(s) as a string in lexicographical order. Each keyword should be wrapped in single quotes and separated by commas.

Added in this fork

Currently only Ruby missing the random tests,

Added in this fork

there are now sample tests in Ruby and the assertions use RSpec. the only thing missing is random tests

Random tests added in this fork

Ruby had this issue, fixed in this fork

Fixed in this fork

0's don't make sense, how should those be treated ..