Platform-computed visibility verdict with the decision evidence in checks (decidedBy, reason, layer statuses, deprecation, platform overrides, and the per-side listing result that folds in curator flags and deprecation). Curated inputs come from the latest published label version; reasons that would reveal a draft claim are reported as their published equivalent. Platform operators read the verdict as evaluated at .../visibility/platform.

The verdict is a labels answer, so it is served for any chain that has labels: on a chain this deployment cannot index, Layer 1 and Layer 2 report that they could not run (decidedBy: unassessable) and the curated inputs still decide, rather than the read being refused.