I do plan to go through the mutation code in detail at some point, but I've done a little investigation on this. What is happening here is that the probability of certain types of mutations has mutated (via a "Delta Mutation" not in the list above) to the point where the probability is high enough that more than one mutation of that type occurs during reproduction for longer genomes. For these types of mutations, the code walks through each base pair of the offspring at reproduction time, rolls the dice and performs the mutation if the dice say to do so. This is why you see the string of identical mutation types all happening on the same cycle but on monotomically increasing positions in the genome. An ancestor bot reproduced that cycle and several mutations of the same type occurred at that point in time (but at different positions in the genome). All the recent virus infections have lengthened the genome to the point where this is more apparent.
I'm going to resolve this specific bug as By Design since the code appears to be working as intended. If you have suggestions for changes on the way mutations should work, I would love to see those in the Suggestions forum. And as above, I will take a detailed walk through this area of the code at some point. In particular, adding some notion of locality to mutation probability so that different parts of the genome can have different mutation rates is something I have long wanted to add.