No skunk on either night. Raccoons. Five raccoon visits across the two nights, two of them working the bare patch with their heads down, and Finn, out for his late-night trips, at the patch twice.
Thursday night into Friday was the busy one: four raccoon visits, including two (possibly three) raccoons together on and around the patch at 01:06. Wednesday night, with the grub killer a few hours old, had one raccoon passing through and no one at the patch except Finn.
First visitor. A raccoon works the grass at lower left, nose down, then drifts back toward the left end of the patch. Ringed tail and mask are clear. It stays on the grass beside the patch.
Two raccoons arrive and keep at it for at least two minutes, heads down along the edge of the bare soil and the grass to its left. A third shows up briefly in the first half. This is foraging on the patch itself. I did not see a clean frame of one digging, so "foraging at the soil" is as far as the footage goes.
Finn, out for a late-night trip. Dark saddle, pale legs, feathered tail, eyeshine as he turns toward the lens. He crosses the patch beside the hosta and leaves to the left in a blur.
Enters from the bottom left and walks up the lawn, ringed tail toward the camera, then heads off toward the play set. Does not stop at the patch.
Last visit before dawn. A raccoon works the upper edge of the bare soil for about a minute, with eyeshine as it lifts its head.
Finn again, standing at the left side of the hosta for about thirty seconds, then leaving. The camera's spotlight fired at 01:23:14, right after it left, and found an empty lawn.
Walks in from the right edge, crosses the bottom of the frame and leaves. Never goes near the patch. The only raccoon on this night.
Lens close-up at 22:18, then the colour waving test at 22:55. The 01:47 spotlight trigger is a bright streak crossing the lawn, most likely a bat or a large insect. Nothing at the patch.
The overnight pass ran on the camera's low-resolution sub stream (640×480, 4 fps), using frame differencing plus a background-subtraction check for animals that stand still. I reviewed the strongest candidates frame by frame, then pulled the main stream (2560×1920) for the visits above. About 470 weak detections on Thursday night were not individually checked; most are wind, insects at the lens and IR flicker.
A small, slow animal at the far edge could still be missed. Wednesday 20:03 to 20:10, during camera setup, is unusable (the NVR export had broken timestamps). The yard is quiet from about 21:00 on both nights, then picks up after midnight.
Species and behaviour here are my reading of the frames, not a detector's output. Raccoon calls are firm; the Finn identification is from the household, not the footage.