Hi, Le mardi 2 juin 2026, 19:30:10 heure d’été d’Europe centrale techmetx11 via Standards a écrit :
As it's written in the specs, I think the challenge protocol can be abused to leak a user's presence, because any device associated to the target user *MUST* send a message to the entity who requested the challenge, as soon as it receives it.
Indeed, that something to change, it should probably be answered only to people with presence access.
This would violate RFC 6121's guarantees to not leak the user's network availability to an entity who's not authorized to know about it.
So, I'm wondering if we could instead have the identity verification be a secondary optional method of verification, and have something like a signature of the JID (signed by the XID) in the user's PEP node, that the user themself HAVE to re-sign once every so often, like a month or week to ensure it remains valid, so that old hosts cannot impersonate users with stale signatures.
That sounds like a good option. Best, Goffi