Every move gets a forecast before action and a score after reality arrives.
Baseline lift · forecast accuracy · execution fidelity · result confidence · evidence quality · claim honesty.
It asks for evidence, routes a human, or marks the move not claimable.