If you are using additional variable then its contribution to the residual and Jacobian must be added manually inside the kernel called by the polycrestal action. Also if there is a contribution from the polycrestal variables to the additional variable it must be added to the multiphase kernel.
You will find an example for such coupling in the polycrestal action itself. If you go to the action you will find ACGBpoly kernel that can add an extra concentration variable to represent the interaction between bubbles and GBs. the concentration equals 1 in the bubble phase and zero in the solid grains. You can turn on this kernel by specifying in the input file c (the concentration) = whatever you want call your additional variable. However, if you want evolve c considering the effect of the order parameters on its evolution you will have to do the same coupling in the kernels that evolve c.