If RP have already identified the user somehow, RP would include allow credential for the authentication. In case of identified user only has single passkey (which means the transport is internal), and the user tries to authenticate, the allow credential only has one credential entry with internal transport.
In this case, I'm thinking that RP is clearly indicates which credential is expected, then browsers or platforms might directly prompt Face ID or Touch ID without showing modal sheet which introduces additional user steps.
For other cases, I understand why passkey modal sheet is needed, but in this case, it would be better to present Face ID or Touch ID prompt without modal UI.
