The Oxford English Dictionary says of "snuck": "pa. tense and pple. also (orig. and chiefly U.S.) snuck.". It gives an instance of "snuck" from 1887.
So "snuck" is an innovation, though in the US it's older than the 20th century. (Like "dove", from 18th C in the US: the first time I encountered "dove" as the past of "dive" in an American source, I was genuinely confused, wondering where pigeons came into it).
I think it's quite rare for new irregular forms to be created, though I can't point to any reference; what has happened a number of times in the history of the language is that two forms have been in contention, and the strong ('irregular') form has eventually won.
I'm thinking particularly of "dug", where you find "digged" in older texts. (However, the OED says that the past participle was always "dug", and what has happened is that this has ousted the original past. It is unusual, though to find a word with a weak past and strong past participle, so I suspect that both forms were there from the outset).
[Edited: I had misinterpreted the OED entry as saying that "snuck" was the original form.]