What problem does it solve? When multiple fixes built on the same assumption keep failing the same gate, this Skill redirects debugging effort from writing yet another fix to questioning the underlying premise that all the failed fixes shared. ## Core Features & Use Cases - Premise Identification: Write down the single sentence that every failed fix assumed, turning repeated failures into evidence about that assumption. - Actor Census: Run a rerunnable census script that counts which actors hold the imbalance, revealing whether the same few actors are assigned the problematic role on every run. - Asymmetry Removal: Rotate, randomize, or relocate the role instead of adding compensating mechanisms like return paths, shared pools, or periodic rebalances. - Use Case: Imagine a load-balancing bug where three consecutive fixes failed and the same two servers keep receiving most of the traffic. Use this Skill to write down the premise, census the traffic per server, find what assigns those servers that role, and remove the asymmetry. ## Quick Start Ask the AI to apply the attack-the-premise principle to a bug where two or more fixes sharing one assumption have already failed the same test or gate.
