I would argue against your veneration of pure logic. Hume’s skepticism, after all, perfectly logical but the system that he proposes is infinitely worse at prediction than the system he seeks to replace. Likewise, theistic logic is often air-tight within its own realm but since I disagree with its most basic assumption I cannot really call it true (its track-record on prediction also leaves something to be desired, largely because it predicts events outside the normal world of experience).
Our brains are not a de novo product, but rather an accumulation over millions of years of evolution, so while our sense may indeed be false, it is worth noting that our brain grew into our senses and not the other way 'round.
Indeed, modern psychology has demonstrated that the mind is informed by the sense, and in the absence of sensual information, the mind goes haywire. So, while I think everyone recognizes that our sense may not be trustworthy, it is a folly to then place trust in the mind, which is dependent upon those perhaps-not-trustworthy-senses. Indeed, in doing so, one is adding an extra layer (rather than removing one) in terms of approaching useful understanding. This is a problem that goes back to the pre-socratics, but that doesn’t mean that it ought be embraced.
Furthermore, since the brain is limited by the sense, to postulate an unknowable (and it is, definitionally that) world where our senses are indeed false is to postulate a world we can never know, can never model, can never interact with . . . and therefore, never care about.
So, what use does such a line of thinking have?
Philosophy that seeks to divorce the mind from the senses (and in turn, the physical realm) will, invariably, experience the same fate that the mind, itself, does when robbed of sensual experience: meaningless madness.
To me, pure logic seeks to do just that – to create an internally consistent system that exists purely in the mind.
Now, I am not suggesting that we throw the baby out with the bathwater! I like logic quite a lot actually, but only insofar as it can be related to meaningful observations rather than from a priori assumptions. Taking things back to first principles only works if it accurately models reality, so we need to have a fully fleshed-out model of reality before we can even begin considering first principles. Most logicians that I have seen try and flip that one on its head and, naturally, the more expansive their reality becomes the more inaccurately their first principles model it.
In short, my way of logically solving this problem would be to start opening boxes rather than mingling my mind with the ether. Otherwise you fall into tricky situations. After all, it can’t be in box 10 because we know that if it were in box 10, by the time we arrived at it, we wouldn’t be surprised. Having ruled out box 10, we run into the same problem with box 9. After all, if it wasn’t in box 8, it would have to be in box 9. Now, we can follow this chain of thought until we discover that it has to be in box 1. Except that it can’t be in box 1, because we’d know that already! But of course, opening boxes with our minds proves nothing. That is my point.