Top. Right. Bottom. Left.

The object uses four edge-aligned destinations. The framework resolves each anchor against the containing stage rather than baking coordinates into transforms.