>> Both designs work logically and produce the correct results. I'd like to know which architecture is generally preferred in terms of best practices for clarity and maintainability.
As far as I am aware, both are acceptable. If you have two endpoints, you could label them differently which may help in some circumstances.
>> I've noticed that when I use the single End point approach, the playbook diagram sometimes visually suggests that both branches (the “Hit Condition” and the “Else”) were activated, even though only one actually executed.
That is a bug which I have opened for the development team to address. Thank you for reporting it.