Can we really say that the specification is complete? It seems that the function that returns 0 only when the list is empty and the maximum otherwise (i.e. there is no special case about negative integers) obeys the same specification.

I found this problem very interesting (I have never attented a logic class, so I always enjoy this kind of exercise where I can learn a bit more about the topic).
But I must admit I find the exercises too short w.r.t. the length of the "details" section. Would it be possible to reason more about the system (in this kata, or another)? That would allow to better understand the system, and have more fun playing with it.