Instead, the password managers each have their own finicky app-to-app mechanism for transferring passkeys from one password manager to another. (I think all the password managers kinda like that lock in.)
It's a nice simplifying step to talk about password managers here, but in the majority of the cases this won't be handled by a password manager, but rather by the device operating system, and the device manufacturers really like that lock in.
They're also in an especially good position to abuse it, because they now know every service that you authenticate and the webauthn "attestation object" field lets them set up a side channel with those services such that they can sell additional information about you.
Some people will tell you that the attestation object is not used in the consumer passkey system, so there's no way for this abuse to occur, but since it's usually going to be the device manufacturer who controls the password-manager-like component here, and they're the ones who stand to profit most from this kind of abuse, I think we need stronger guarantees than "the spec says you shouldn't do this unless the user is your employee and you paid for their device".
Until they fix this, I'm sticking with my mess of TOTP authenticators and yubikeys.