That’s why I said largely useless. An attacker can narrow down the attack surface by ignoring anything that can’t login, but that just leaves them with root and delial, and they already knew or could’ve guessed both of those pieces of information (in this context anyway).

And as you noted when looking at the service accounts, they might be able to login or crack their way in via xrdp or sshd. So, unless you’re port-forwarding those protocols from the internet, how useful is that really? I would say largely useless. Assuming they port-scanned your public IP, they still need either an insecure config or an unpatched, remotely exploitable bug.

That being said, you’re totally right. The average Linux user isn’t “administering” their system, so they probably aren’t following their distribution’s security mailing list, installing security patches as they’re released, and actually RTFM. It’s best for the average user to play it unbelievably safe.

In this case, the machine isn’t actually running xrdp, and sshd doesn’t accept passwords or root logins. (Although, I need to setup knockd to protect that non-standard sshd port a bit more.) All passwords used on the system are random and longer than 32 characters. My router doesn’t port-forward to this machine, either.

This has been an exercise of Cunningham’s Law for the benefit of those reading.