Post #2501069

2026-05-07 23:07 UTC

I&#39;ve been wondering ... with the current drive to increase #Passkey adoption (which I support), as well as with the acceleration of #PostQuantum cipher adoption (often with a target of 2029), I believe there&#39;s a problem emerging. Passkeys/WebAuthn take some time to adopt, and there&#39;s a lot of organisational effort involved to get all the ducks in a row. As there doesn&#39;t seem to be a post-quantum resistant FIDO standard in place, yet (please, correct me if I&#39;m wrong), they **would have to go about it again.** There does not seem to be any cryptographic agility in the system at present. Particularly with PQC hardware tokens (Yubikey, @nitrokey and others) unavailable at present, this will become even more difficult. How do others see this? I&#39;d love to have a bit more knowledge on long term sustainability of the efforts in migrating on a broader basis towards Passkeys. #cryptography #PQC

Replies (2)

@phlogiston@mastodon.nz 2026-05-07 23:24

There seems to be an ML-KEM WebAuthn draft: https://datatracker.ietf.org/doc/draft-vitap-ml-dsa-webauthn/ And the 'advanced' section of this FIDO2/WebAuthn PQC Testing Platform and Dev Tools are suggesting ML-KEM testability. https://webauthnlab.tech/ But it all seems still a bit 'further out there' to give me warm fuzzies without having to face another disruptive transformation yet again.

Open ##2635071
@iinuwa@fosstodon.org 2026-05-08 16:02

@phlogiston@mastodon.nz @nitrokey@social.nitrokey.com ML-DSA is supported by the protocol today via cryptographic agility, but RPs and authenticators need to catch up. If RPs update their servers to support and prefer ML-DSA then, whenever authenticators support ML-DSA, they will register PQ-safe keys. There is a separate issue of how to support upgrading authenticators that support both classical and PQ algorithms and have previously created a classical credential to use a new PQ cred. More discussion is needed there.

Open ##2635072