Analysis and attempts to decode the Voynich manuscript lead me to believe mathematical patterns intended to hide information, languages in particular, are not safe in the least.
When the work required to brute-force a cipher in a sane timeframe doesn't exist (or even couldn't fit) within the bounds of the observable universe, (if the crypto works as intended, which, to be fair, is a big if) brute-forcing is safely axed as an avenue of attack.
Of course, on these scales, if you really want to, you could connive and threaten your way into having a backdoor installed. Or you could spy on the victim and steal their laptop, with the key decrypted and in memory. Or you could beat the key and password out of them. Brute-forcing shouldn't be the most pressing of anyone's cryptographic worries.
When software uses a known set of hard coded primes, it is logarithmically simplified in what it takes to defeat it.
Maybe there are good primes, bad primes, but to "standardize" a prime is to give an anchor to an adversary.
The trade floors are already opening in Asia, I think. You have got 24 hours. Go!!!