To experiment, I built a default wing and increased spanwise resolution -- Then I turned off symmetry and copy/pasted a duplicate and rotated that duplicate 180 about X.
I then ran it at 10 deg alpha -- no propellers.
When generating the contours in viewer, it clearly shows that the sign convention for the two wings is opposite -- one shows large negative delta Cp and the other large positive.
I have two other suggestions that may work for your situation -- they're a bit of a hassle, but may be worth it for you. In particular it depends on how many (and what kinds) of variables you want to vary in your design process.
When we started the discussion, it sounded like there was going to be a lot of focus on planform variables (strong desire to link the left and right wing halves, etc). Now it seems like there is more focus on the twist variables.
1) You could build your model with a single wing component -- starting at left wingtip, to center, to right wingtip -- but without any symmetry. You'll have 'inverted' taper on the left wing panel, then normal taper on the right wing panel. This will keep all the parts oriented properly and also give you separate control of twist on both halves of the wing.
2) You could use a single bladed propeller to represent your wing. While this is a bit of a pain, it will actually give you the greatest amount of control of camber or twist distribution -- you could have a simple straight planform with a highly nonlinear twist distribution.
I've attached a simple example that includes both approaches.
Rob