MSC3879: Trusted key forwards #3879
Conversation
uhoreg
added
e2e
proposal
Co-authored-by: Denis Kasak <dkasak@termina.org.uk>
| that the events encrypted with this session cannot be trusted, which causes | ||
| confusion for users. | ||
|
|
||
| We propose adding a flag to the `m.forwarded_room_key` event to indicate |
There was a problem hiding this comment.
Maybe clarify that it's added to the event content
| @@ -0,0 +1,75 @@ | |||
| # MSC3879: Trusted key forwards | |||
There was a problem hiding this comment.
As mentioned elsewhere, the term "trusted" isn't a great term to use. It is used to mean various things in various places. We'll probably switch to another term.
The basic premise of this MSC is to provide a method for the key forwarder to indicate whether it believes that the room key was actually sent by the given sender_key. It is still up to the recipient of the key forward to determine whether the sender_key is "trusted" (e.g. by checking whether they have been verified)
There was a problem hiding this comment.
As mentioned elsewhere, the term "trusted" isn't a great term to use. It is used to mean various things in various places. We'll probably switch to another term.
My objection to "trusted" isn't particularly about the term itself, but rather that it tends to get used without any additional context (what does it mean for a device to be "trusted"?). Provided it's clear from the context what we actually mean by "trusted", it's fine.
There was a problem hiding this comment.
I think that "authenticated" would be a better term than "trusted". I used "authenticated" for MSC4048, and I think we should use it here too. In either case, both should use the same name.
Rendered