Staff chooses one full held source/own four questions after teaching/suitability/exposure review. Keep other stages, ordinary solved models, keys and all-stage data unseen before capture. Usual communication/private untimed/wanted reading/scribing/rest supported; mathematical hints distinguish access. Record actual initial versus supported response, exposure/support/component/teaching/interval separately. Uncaptured evidence is unavailable. Later/reserve is not automatic numerical growth or a secure-equated test.