No, that’s interesting. There is also TLS 1.3 and NOISE/WireGuard channel messaging doing similar at the AEAD level (and original age, of course). If you just look at the diff between the original age spec and this proposal, you don’t really have to argue at the ground level. But since it’s cryptography, one ought to be suspicious.
There are plenty of libraries with age support out there, which can be made compatible in a breeze, and we inherit all the nice header unlock features like SSH pubkey or Yubikey, without changes. Simplicity is key.
No, that’s interesting. There is also TLS 1.3 and NOISE/WireGuard channel messaging doing similar at the AEAD level (and original age, of course). If you just look at the diff between the original age spec and this proposal, you don’t really have to argue at the ground level. But since it’s cryptography, one ought to be suspicious.
There are plenty of libraries with age support out there, which can be made compatible in a breeze, and we inherit all the nice header unlock features like SSH pubkey or Yubikey, without changes. Simplicity is key.