Hypothetically, science proves free will isn't real

[Angrboda]

… it is not possible for reason to overcome its limits no matter what your existential commitment…there is no way to tell if we are or are not at the last layer of the onion.  It could go on further for all we know.  Your commitments don't resolve that quandary.  Reason, ultimately, cannot provide justification for itself.

Not sure if we are debating whether human beings imperfectly apply reason or whether reason itself is faulty.

Not sure what you see as the difference. I would argue that reason is both reliable and faulty. It is reliable in that it can produce consistent answers within its domain. It is faulty in that cognitive biases and irrational impulses tend to determine the conclusions of reason more than do those aspects within our conscious, reasoning apprehension.

I also believe that first-principles are truly self-evident. You seem to be suggesting that they may only appear to be self-evident. If first principles are up for grabs then pretty much everything is. Do you disagree?

They default to self-evident because they cannot be justified, but our intuition tells us they are true. That isn't in the bailiwick of reason. They may appear self-evident, but that tells us little about their connection to reality. The assumptions behind local realism appear self evident, nonetheless the postulates of local realism are at odds with quantum mechanical experiments, so one of them is wrong. Self-evidence is no guard against error.

… your argument that these paradoxes and seeming absurdities are a problem is at heart an argument from ignorance…

Paradoxes serve as indications that something is wrong with the theories that produce them. In the absence of a plausible theory, one that doesn’t resolve the paradox, more inquiry is required. But that’s not what’s happening here. These are not questions that yield to empirical findings. Some lines of reasoning are true dead ends. Moderate realism avoids paradox whereas the known alternatives do not.  It makes no sense to reject a plausible theory when there is no good reason to oppose it, just because you think someday, maybe an undiscovered alternative will appear to redeem your doubt.

From what I've seen of moderate realism, it doesn't resolve the paradoxes so much as covers them over. In that it is no less a flawed theory than alternatives. Regardless, it doesn't change the fact that your line of argument is based on ignorance, and therefore your conclusions are unreliable.