A good test question there (I realize I was wrong about the (A, B, C) example. You have to knock out all three.
You have two contradictory hypotheses: "if claim 1 was invalidated by a prior art fly-swatter" and "If claim 2 was invalidated by a prior art fly-swatter"
If fly-swatters are prior art (as they are), then you would claim your robot thingie as :
"An automated fly destroyer, comprising:
a fly swatting device,
and a robot configured to etc. etc."
And you would say in your spec, "the fly-swatting device can include, in one embodiment, a commercial fly-swatter, but could be any device capable of killing a fly, including newspapers, rolled up magazine, board, book, and many more."
However, if your claim enumerated the features of a fly-swatter, then it would certainly be questioned. You should not bother listing things that are known in the art, and furthermore, you'd have failed to claim the book, newspaper, magazine, etc. methods of killing the fly.
On further editing: your idea of knocking out a dependent claim by knocking out all the elements is theoretically correct. However, that's not the way things are usually done.
[1] explains the issues around this. IPRs almost always go after the broadest claims, and the patentee needs to respond "even if claim 1 is invalidated, claim 2 is still valid because blah blah blah."
Patents almost always put the heavy novelty in claim 1, like I did for your robot fly-swatter invention. The dependent claims are narrower because they add additional elements which often aren't very novel. However, you're right that they can still survive and be useful.
[1] https://www.natlawreview.com/article/patent-owner-tip-12-sur...