edge-case-challengerlisted

# Edge Case Challenger A pressure-test skill for a decision that already works. It surfaces the question a working solution stopped anyone from asking, by running a fixed set of probes against that one decision. The value is DISCIPLINE, not paranoia. "What could go wrong?" free-associated gives a generic risk list. Taking each probe and forcing it onto the specific decision, including the probes that indict a choice the user is proud of, gives failure modes bolted to their reality. Two named mechanics are why a working decision resists challenge, and the skill exists to defeat both: - **Einstellung effect** (Luchins 1942; Bilalić/McLeod/Gobet, *Cognition* 2008): a solution that works blocks the search for a question about it. In the chess study, experts who had found one move *reported* they were looking for a better one, but their eyes never left the features of the move they already had. The lesson that shapes this skill: self-report is worthless. Asking the user "did you reconsider?" gets a yes while nothing actually moves. Only a concrete probe moves attention. - **Path dependence / absorbed cost:** an early decision makes some quantity part of the terrain, then a *second* decision removes the pain of that quantity, so nothing ever forces a re-look. The pain, not the wrongness, is what would have triggered review; kill the pain and a questionable premise lives untouched for months. Sibling skills: step-back-as-maintainer (re-enter code as its future reader) and break