Thanks Tom,
I am not an engineer, but it seems that it is necessary to define a bunch of specifications for these 3D bodies.
As Pilou stated, the densities, and 3D depth are two parameters.

It is pretty easy to get the centerpoint of a uniform "cuboid".
What is meant by "Shape"? Are the "cuboids" to be bent?, have portions removed?
Touch each other or be separated?
The combined center of gravity would be located back in the 3rd dimension.
If mass A is fixed in relation to the combined center of gravity, mass B looks to be a bit low.
It seems possible that with too many constraints, the combined center of gravity will be elsewhere than where desired...
- Brian

Pretty sure that the three centers of gravity need to be in a straight line...