What problem does it solve? When two or more fixes built on the same assumption keep failing the same gate, teams often write yet another fix instead of questioning the underlying premise. This Skill interrupts that loop by forcing the premise to be written down and by taking a census of which actors hold the imbalance before any new fix is attempted. ## Core Features & Use Cases - Premise Extraction: Writes down the single sentence every failed fix assumed, turning an implicit belief into a testable claim. - Actor Imbalance Census: Counts which actors hold the imbalance per run, implemented as a rerunnable script, to reveal who is repeatedly assigned the problematic role. - Asymmetry Removal: Rotates, randomizes, or relocates the role instead of adding compensating work like return paths or periodic rebalances. - Use Case: A queue consumer keeps overloading the same worker despite three retry and throttling fixes. Apply this Skill to write down the premise, census the load per worker, discover the scheduler assigns the role, and rotate the assignment instead of adding another patch. ## Quick Start Ask the agent to apply the attack-the-premise principle to your repeated failure, providing the failed fixes, the shared gate, and the decision this analysis should change.
