Dead author ethical guidelines
(My personal take on this matter - which is only based on intuition/opinion rather than experience):
When you're about to publish such a paper, ask yourself: "How certain am I that the deceased would have put their name on the paper?"
If the answer is "certain", go ahead and name them a co-author. But - explain that the attribution is posthumous, both in a footnote from the person's name (the same kind of footnote used for affiliation) and in one of the opening lines (perhaps a note as the first line.)
If the answer is "not certain", but the deceased's contribution was significant, don't make them a coauthor, but devote a few sentences to emphasize the significance of their contribution or their involvement, and clarify that the work had progressed or diverged beyond what you had worked on with the deceased (or some such explanation).
If the answer is "would probably not", and the deceased's contribution was not significant enough for co-authorship, then you just mention them in a thank-you note somewhere in the paper, same as if they were alive.
As for "preventing dishonesty" - if you're an editor or non-blind reviewer, and notice the name of a deceased person on a paper, without an explanation of the deceased's contribution within the paper - reach out and ask the corresponding author for a side-note (outside the paper) explaining the deceased's involvement with the submitted work.
I don't see any dilemma here. If a person would be a co-author if alive, then they should still be a co-author if they die before submission. But if they contributed less and so wouldn't be a co-author if alive, then it may be appropriate to acknowledge their contribution in an appropriate section or footnote.
Even if I disagree with your conclusions in a paper, it may be appropriate to ack me, even if you don't mention the disagreement. But that is a judgement call you need to make.
I don't have a general solution for the last part of your question (name-dropping). I worry that it might require an investigation into the provenance of a paper that we seldom do as a matter of course. But it would be pretty obvious in some cases. I could try to publish a paper with Paul Erdős, for example, but it would be immediately recognized as a scam.