Right, if you phish the credentials to an account that is not protected by a passkey, one thing you can now do is set up your own passkey on it. That's exactly what passkeys would've protected the victim against!

The other articles you've linked also describe attacks against accounts that allow non-passkey logins, or attacks that require a compromised browser.

There will never be a login that doesn't involve a potentially compromised device/OS/browser/password manager. As long as using a passkey means using those things too, having a passkey isn't enough to keep attackers out of your accounts. They can be pretty good about keeping legitimate users out of their accounts though. You just have to weigh that risk against the risk of you falling for a phishing attack.