Here's the solution that I used to shoot the beast at https://www.photomacrography.net/forum/ ... hp?t=40624 .

The main component consists of a small piece of balsa, with three short segments of insect pin embedded in it, epoxied to the middle of the bug's back. For photography, the pins stick onto a small magnet which in turn is epoxied to the end of a piece of balsa, which can then be pinned in any orientation needed for photography.
The image above shows my prototype, attached to a dead and dried specimen.
Here's the setup in action, holding the live bug on its back so that I could photograph the underside of the thorax.

Here are the images seen by the camera with 2X and 20X objectives:


The cuticle of the bug is a little bit flexible, so at 20X it moves around quite noticeably when the bug waves its legs, stretches its neck, and so on. But I've concentrated on an area that's close to the glued fitting, and if I'm lucky, the movement will be mostly correctable by Zerene Stacker's alignment procedure (X,Y,Rotate), leading to a clean final result.
Here are animated GIFs showing the worst two adjacent frames in the 20X stack, before and after alignment.


When the photo session is over, the "backpack" just pulls away from the magnet, and then what's left is light enough that the bug can move around pretty much unencumbered. Kept in my usual plastic boxes, the bug has no trouble climbing vertical walls. But interestingly, the bug is now not willing to attempt hanging from the top of the box, though it regularly did that before I added the fitting. Occasionally it puts a front foot on the top, but that's as far as it goes. All I can guess is that somehow it figures out how much grip it will have, decides that it's not enough, and pulls back. This is actually quite a relief for me, because I was having bad visions of the critter trying and falling, again and again. No problems there, though.

--Rik