My understanding is that this is a key tenant of visionOS’s design, where apps don’t get access to gaze data (I think unless they’re taking over the full screen?)