[Jeremy]

It's not something I'm aware anyone else has done before.

But it's still something I think is achievable if you are ready for a bit of Python programming.

I wouldn't go the onAction route.
What I would do instead is "Modifying the program" once the 3D printing project is done in RoboDK but before post-processing it.
In other words, I would modify the "white" program generated. (You can "right-click" -> "Show instructions" on the white program to see the instructions in it.)

Here's an example of how to modify the instructions in a program://www.sinclairbody.com/doc/en/PythonAPI/exam...structions
(You might want to try it out on a simple program you create manually (like 4 targets, 4 moves) before trying on a 3D printing program.)

My strategy would be to find the first relevant move, find the second relevant move, figure out the "angle or orientation" in-between the two, and add a program call to reorient the axis in the right direction in-between the 2 points. Rince and repeat in-between each point.
For the simulation, you can use a "start thread" instead of a "program call" as it's non-blocking.
For the real robot, it will depend on how you manage the external axis.

I wouldn't say it's an easy solution, but I think it shouldn't be too hard either.

Hope it helps.
Jeremy