I think this question really has two parts, the ethical question and the practical question.
First of, the ethical question. Let me say that I am surprised that you see the issue whether it is ethical to self-plug your work on Wikipedia as an entirely different issue than whether this is allowed by Wikipedia policy. Wikipedia is a private web site run on private donations, and is entitled to define how they want to do things around their piece of the web (much like we do here at Stack Exchange, to the chagrin of plenty a new user). I think it is impossible to argue that it is ethical (for academics or anybody else) to (mis-)use their service in a way that violates their own rules for your personal gain. Hence, to answer your updated question:
It's a question about what is considered ethical in academia
This is not, precisely due to the reason that you say you are not interested in: it is not allowed by Wikipedia. It cannot be ethical to use their service, but don't consider their rules.
Second, your question also has a bit of a practical part. You claim that it should be really easy to self-plug your work on Wikipedia, and that it will be hard to trace and prevent. (You also seem to imply that this impact on the ethical question, which I think it has not - whether something can be done is pretty much irrelevant to whether it is ethical to do it)
I have a bit of a case study for this. A large European research project I was involved with some time ago once got asked by its reviewers to add their own papers and other outcomes to relevant articles "to increase visibility". After some internal discussion, the project obliged and started with some really defendable and arguably independently useful edits (taking very stubby articles and improving them with peer-reviewed content from our own papers, and adding references). All of those changes were rolled back within minutes, with the explanation that the edits looked too much like self-advertisement. We tried again, even more conservatively, and the same thing happened again. It turned out that the self-policing in Wikipedia works much better than anybody in the project anticipated. We gave up on the issue quickly.
The morale here is just that Wikipedia is aware that people could try to do this (and I am sure many do), and are very vigilant. I am sure you are able to edit in your links given enough energy, but it may be much harder than you imply it is.