Yesterday was less humid in the house, and it took an extra puff
to trigger the lead error, but it was still easy. Maybe the shroud
of the jack uses a water-absorbing plastic like nylon.

Maybe there's some hygroscopic shmutz around the jacks. Have
you thoroughly cleaned/flushed the area?

Yeah, it's clean, plus it's basically new. I've not been dragging it
around a steel mill or anything like that.

This has puzzled me ever since you first posted.

CMOS is sensitive to surface contamination -- anything that's conductive,
even only slightly including water.

You might try hitting the area with freeze spray. This ought to reveal two
things... First, is the problem caused simply by blowing on the board (ie,
moisture has nothing to do with it). Second, the low temperature will cause
moisture to condense. Does the problem occur after the spray has evaporated?

If I had access to an ultrasound bath, I'd dunk the board in an inert
solvent and clean it thoroughly. If that didn't clear the problem, I'd be
surprised.