The circuit board is 30-mil, double-sided, half-ounce (I think) copper on glass-fiber stock, direct-etched by rubbing ferric chloride solution onto it with a sponge.
Used the CNC Sherline to drill the holes; the G-Code is now tailored for my Sherline mill and tool-length probe station.
The copper layers as a 600 dpi PNG file:

The top copper image (on the left) is reversed so it comes out correctly when you’re doing toner-transfer etching.
I didn’t bother with a silkscreen, because I don’t have a soldermask and there’s no room for text around the parts anyway.
The four vias at the corners mark the edge of the board. Trim it with tinsnips (or a shear if you have one), then introduce it to Mr Belt Sander until the edges pass directly through the middle of those via holes. Round the corners a bit so they fit into the case recess atop the mounting shoulder.
Put Z-wires in the small round vias (the ones that don’t have any other traces) to connect the top and bottom ground planes.
Put Z-wires in the other round vias to connect a top-side signal to the corresponding bottom-side trace.
There are three jumper wires across the bottom; with only two layers I don’t get all bothered about embedding the last few. Those vias are square.
I don’t have any way to do plated-through holes, so solder the wires to both sides of any vias with traces on both planes. I admit I missed two of them on the TT3 ribbon cable.
The big empty space around the positive power terminal prevents the ring-lug connector from shorting to the ground plane. Now that I think of it, there’s no need for an empty space on the bottom copper, but it doesn’t do any harm.