[gruez]

>You may want to enroll multiple and keep them in other places too, but you can't enroll a key you don't have so things like a safe deposit box are not useful for the average case.

That seems like a usability nightmare. Are there plans to improve this? Hardware wallets for cryptocurrencies seem to have it solved. You can keep multiple copies of the keys around (ie. multisig wallets) for maximum security, or you can write down the private key of the device you have and store it somewhere safe. In either case you can retain the public keys so you don't need access to the device if you want to send funds to them (or in the case of authentication tokens, enroll them).

[vorpalhex]

Because each hardware key is unique, this is not a feature currently available nor likely to become available. Each token from the yubikey is not (readily) linkable to the key itself since the underlying secret is opaque and can't be exported, so tricks like Shamir's aren't readily possible.

Yubikeys do solve a lot of use cases very well but that is a downside to them. That is probably still a good tradeoff for most consumers.

[gruez]

>Because each hardware key is unique, this is not a feature currently available nor likely to become available.

You don't necessarily have to do it crypto wallet style and have the private key be exportable. Just adding a public key export (on the security token side) and a way to enroll a token by its public key (on the browser/website side) would allow you to enable 2fa without having to make a trip to the safe deposit box (either to store your backup codes, or to fetch your backup token for enrollment).

>Each token from the yubikey is not (readily) linkable to the key itself since the underlying secret is opaque and can't be exported

That's not an issue. You can derive more ECDSA public keys from a single master ECDSA public key[1]. The corresponding private keys can only be derived using the corresponding master ECDSA private key, and the generated public keys can't be linked back to the master ECDCSA public key. Bitcoin hierarchical deterministic uses this property to generate wallets that don't need regular backup (all your addresses are derived from one key) and apple's find my network uses something similar.

[1] exact mechanism is described here: https://bitcointalk.org/index.php?topic=19137.msg239768#msg2... starting at "Type-2 is a bit less obvious [...]"

[tialaramex]

For FIDO (and thus WebAuthn, and thus to make this actually practical beyond a toy that only works for some particular Yubico product) the keys are random per enrollment. This is intentional because it means that you can't be tracked, since "your" key on Facebook and "your" key on GitHub are no more related to each other than "my" key on Facebook is to "your" key on GitHub.

Google have apparently some plans to address this problem in the medium term. Adam Langley has written vaguely on this subject before. In the short term, their priority is the trick he wrote about most recently - if your Android phone is enrolled as a Security Key with Google, and it's signed in to Google because it's an Android phone, and you use Chrome on a desktop, which is also signed into Google, the Chrome can use Bluetooth to determine if the phone is physically nearby and if so propose to authenticate your desktop Chrome to a remote web site using the Android phone. Elegant, albeit not suitable for those who fear lock-in.

[gruez]

>This is intentional because it means that you can't be tracked, since "your" key on Facebook and "your" key on GitHub are no more related to each other than "my" key on Facebook is to "your" key on GitHub.

I get the motivation behind it, but the mechanism I proposed in the last comment still preserves those properties? Each site would still get its own derived ECDSA public key. The master ECDSA public key would only be shown to the user and is to be kept within the browser. If a user wants to enroll a not-present security token, the browser will take the ECDSA public key and derive a public key to present to the site, so the site still can't track users using security tokens.

[tialaramex]

To complete enrollment you need to know the corresponding private key, live.

The relying party says "I am some.example and I want to enroll a Security Key, but, not ones which recognise these huge random-looking IDs that are already enrolled: 12345678, 34561234. I also picked this random nonsense XYZXYZXYZ. Go for it" and your browser talks to your Security Keys until it finds one that isn't already enrolled, gets that one to sign the appropriate message and sends back, "I am a web browser, I checked that you are some.example. I picked my own random nonsense XXXZZZ, and a Security Key picked public key ABCDEF, then to prove it knows the private key it signed this message for some.example mentioning XYZXYZXYZ and XXXZZZ and with bitflags it understands enough to know what it's signing. It says the resulting credentials have random-looking ID 98765431. Thanks". /s/Security Key

If the Security Key cost less than a low-end Yubikey, it has no storage. That random-looking ID is in some sense your private key for the site, but suitably encrypted, e.g. with AES in Galois/Counter Mode, so that the device needn't remember it, when a site asks keys to authenticate, it must provide the ID they're authenticating against, and so they can do AES GCM, figure out if they minted this ID, and if so recover the private key and authenticate. This is fiendishly clever, but so far as I can see renders your idea impossible.

[gruez]

>you need to know the corresponding private key, live.

This seems like the main blocker. Why is that required? In theory all the site needs is a public key to verify against.

[vorpalhex]

It looks like Yubikey supports ECDSA keys as of 5.2.3 (Yubikey 5+ devices) and will export the public keys and allows private key signing so this should be possible. It will be an irregular yubikey flow code wise but user wise will appear normal.