The IETF (or whomever) could pre-specify a nothing up my sleeve number algorithm. For example they could choose Pi and say that the following are the official

The IETF (or whomever) could pre-specify a nothing up my sleeve number algorithm. For example they could choose Pi and say that the following are the official NUMSN for each order of magnitude: 1, 41, 592, 6535, 89793, 238462, 6433832, ... and if a number is needed with a specific property (e.g. 40 digits and prime) than the nearest number to a NUMSN is the official NUMSN for that purpose. Then any future standard that used a constant they claimed was a NUMSN but wasn't complaint with RFC 31415: Procedures of Generating Constants would be suspicious.