Light display text on the lacquer gradient

This paragraph sits directly on the body's dark gradient ground. Its near-white ink clears WCAG AA against both gradient stops by a wide margin, so a detector that measures the real ground must not flag it. A detector that assumes a white body flags it at about 1.3:1.

Light text on a raised solid oklch surface between the text and the body gradient still resolves through the normal opaque-ancestor walk.

Dark ink on a translucent light wash over the dark gradient. The wash composites to a light surface, so this passes; measuring raw gradient stops would wrongly flag it.

Light ink on the same translucent light wash genuinely fails once the wash is composited over the stops; raw stops would wrongly pass it.

This muted dark-gray ink genuinely fails contrast against both stops of the dark gradient ground, which proves the gradient stops are being measured instead of the checks silently skipping.

Muted ink reached through a transparent wrapper still resolves to the body gradient and still fails against its stops.

A non-body gradient with legacy hex stops keeps working through the original rgb/hex parser, and light text on it passes.

Dark ink on a translucent wash over a texture image. The image pixels are unknowable to the analytic walk, so contrast must skip rather than composite the wash over the wrong base.

Dark ink on a photo-backed panel sitting over the dark body gradient. Walking past the photo to the gradient stops would wrongly flag this; the unmeasurable image layer must end the walk.

Muted ink inside an opacity-zero scene deck. Nobody sees this state, so the contrast checks must not measure it.