You reckon? I don't think I'm nearly as confident: genuine proofs of security are pretty much unheard of except the kind that prove an attack is at least as hard as some other problem which hasn't actually been proved infeasible. I wouldn't be terribly surprised to find that the shortest proof of the unbreakability of the crypto algorithm consisted of a counterexample to every purported attack on it, and hence was infinitely long.

Still, it'd be worth a try, I suppose.

Well, you might need to prove P/=NP along the way, but surely that wouldn't be much of a problem? If it turned out that P=NP there'd be more of an issue with making a secure crypto algorithm (and thus with proving one), I guess.

Of course, in either case you could gain credibility with your Fermat proof, which would probably be enough to get someone to take a serious look at anything else you did.