In order to act on a claim (inaction is also an action; "avoid committing oneself to accepting" is just ornate wording for choosing inaction) one must evaluate the "probability of truth" of said claim. To my knowledge, scientific consensus is the most reliable source to inform that evaluation. Since you fail to provide a more effective alternative, then it seems that scientific consensus is at least non-inferior to any other method either of us is currently aware of. Thus it is the rational choice on which to base action; basing them on a less reliable source gives a higher probability of being "wrong", and thus would be irrational.
> I see no evidence that this narrative is generally true.
Honest answer: Then I invite you to attend a consensus conference, or enough of them for the p-value on your evaluation of how they run to reach whatever level is acceptable to you. Tongue-in-cheek answer: You also have no evidence you aren't a simulated agent in an advanced species' super-computer (Descartes's "brain in a jar"). At some point, you have to choose some axioms and build your world-view from there.