I notice that, despite getting better at this, I still have a worrying tendency to uncritically accept novel scientific hypotheses without sufficiently digging into their experimental support. I'm guessing that part of this is because schooling mostly teaches us to just unconditionally accept whatever is written in our textbooks as the truth. (It does get much better in university, but there too it could still be considerably improved.)

That would suggest that tests in school should be less "you were taught a theory in class, now explain everything about it" and more "you were taught three contrasting theories in class, now compare their plausibility based on their strengths and weaknesses that were discussed".

Probably the extent to which kids were taught theories-as-facts vs. many-contrasting-theories should depend on the extent to which we did know things for certain. E.g. it could be appropriate to teach physics mostly as facts, because we really do have a lot of physics quite nailed down and knowing a lot of physics facts helps recognize many forms of fraud and crackpottery as exactly that. On the other hand, in subjects where there's a lot of uncertainty, it may better to teach critical evaluation of those subjects than theories which might go out of date within some decades anyway.