You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
(I'm using the "root v11" label to make sure this gets discussed but I don't think we need to rush with this.)
I believe ysa-2024-03 affects some or all yubikeys used in sigstore root-signing. An attacker could duplicate elliptic curve signing keys on these yubikeys. The factors that makes this less severe are
attacker needs the PIN
attacker needs physical possession of the yubikey
attacker needs specialized equipment
My opinion is that we should phase out current keys but that it is not critical to do it right now. Potential fixes that can be done during a root signing event:
Switch to a non-affected algorithm (ed25519 or RSA) -- knowing that this could affect client compatibility
Switch to yubikeys with firmware >= 5.7.0 -- this seems like the better choice
Issues to keep in mind:
tuf-on-ci root key rotation may need a bit of work if threshold of keys change at once (but the signer identities remain same): this is a tricky case where signatures from both old and new keys are required test root key rotation when threshold of keys rotate theupdateframework/tuf-on-ci#505 and likely has not been fully implemented for this specific case
An alternative may be to only change less than threshold keys at a time
The text was updated successfully, but these errors were encountered:
Agreed that this does not seem critical to address immediately, though if possible it would be nice to distribute updated Yubikeys before the signing event. Should we ping keyholders to see if they have access to a newer key, or ask them now to obtain one?
I'd prefer to continue to use ECDSA. I recall some issues supporting Ed25519 for the Ruby ecosystem, and RSA significantly increases the size of metadata.
Does tuf-on-ci check firmware versions (I assume from device attestations) for Yubikeys, or is this a feature we should add? Do you think there is value in copying device attestations into either the repo or the root.json file?
jku
changed the title
[root v11] yubikey updates
[root v12] yubikey updates
Feb 3, 2025
(I'm using the "root v11" label to make sure this gets discussed but I don't think we need to rush with this.)
I believe ysa-2024-03 affects some or all yubikeys used in sigstore root-signing. An attacker could duplicate elliptic curve signing keys on these yubikeys. The factors that makes this less severe are
My opinion is that we should phase out current keys but that it is not critical to do it right now. Potential fixes that can be done during a root signing event:
Issues to keep in mind:
The text was updated successfully, but these errors were encountered: