Deciding Under Uncertainty

The failure this fixes: asked a consequential "should we?", an agent researches well, then lands the answer as a verdict - "My call: X. That's a weekend." Confident, plausible, and missing everything that makes a decision defensible or cheap to exit: no odds, no reference class, no failure rehearsal, no pre-agreed way out. Advice on a hard-to-reverse call is a bet; deliver it with a bet's machinery.

The classic checks (sunk cost, pros/cons, the successor test) need no help - don't pad the answer with them. The five moves below are the ones consistently skipped.

0 - Gate

Same doors as design-forking. Cheap to reverse: answer plainly, note the reversal path in one line, stop - odds and tripwires on a two-way door are ceremony. Hard to reverse with only one option on the table: widen the set first (design-forking owns that), then return here to commit to the winner.