Focusing and a technical "felt sense"

I started reading Focusing a few weeks back after stumbling into self-help (or do we call it "lifestyle design" now?) twitter. The book offers the idea of a "felt sense" of problems where the body knows something is off, separate from and unaffected by any mental or verbal thoughts you might have.

An example: you've just left the house and get the nagging sense that something is off. After a bit, "Hm I think I forgot something at home..." From there: "Is it my drivers license? Headphones? Costco card? ..." When you finally remember, there's a palpable sense of relief, a shift where something releases, like the body saying "yep, that's it." The important part is that a niggling feeling came before the thought, and no amount of verbal brute forcing makes it go away until you actually answer the question behind the feeling. Focusing argues you can use this process as the hammer to many of life's nails: find the felt sense of a problem and use the "body shift" as the guide to judge solutions.

Self-help woo aside, I realized that I have very similar feelings... when looking at code. Especially during code review, I'll scan a piece of code that nominally works but something "feels off." That "off" reaction might appear in an instant, but then I need the next couple minutes to unpack and figure out exactly what it is. Maybe it's a bug, or a performance issue, or a structure problem that will be painful to fix later. But when my brain has finished working it out, there is the feeling of "ahh yeah, that was it."

A couple followup questions:

• Do experienced professionals in other fields do this? e.g. does a doctor have a felt sense of a patient needing the emergency room, or an architect the felt sense of a room's structure, or an investor the felt sense of a company's P&L "health"?
• Assuming this felt sense is A Thing and a marker of some expertise, what's the best way to build it up? Is it possible to build up in all domains? Can we identify it in others as a marker of real expertise vs surface-level learning, i.e. the Max Plancks vs the chauffeurs?