Re: [webauthn] §6.1.1. Signature Counter Considerations does not explicitly mention constant-zero case (#1734)

> @MasterKale Are you saying that signature counters are going away? That would be a pity because they add one element to secure replay detection when FIDO is used for payment authorizations. This is a core feature of EMV (secure payment cards).
> 
> Android still supports signature counters.

Single-device bound credentials tend to still have counters, IE yubikeys, or a touchid from apple that is attested and device bound. However passkeys do NOT have the ability to synchronise their counters in real time and so they have counters always == 0. 

So IMO there are two states we should have in the standard.

At registration you register a counter of 0, indicating you will never use a counter with this device and it will always be 0 (and likely a passkey ....).

At registration you register with a count of non-zero, meaning the counter will increment as authentications proceed and the RP should update the counter state over time. The counter will always be non-zero.

In the states of registration is counter == 0 and auth has non-zero counter, or counter is non zero at registration and during authentication it goes to 0, both should be RP security choices (recommending to reject). 

-- 
GitHub Notification of comment by Firstyear
Please view or discuss this issue at https://github.com/w3c/webauthn/issues/1734#issuecomment-1140550429 using your GitHub account


-- 
Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config

Received on Sunday, 29 May 2022 23:59:09 UTC