Departures from standard spherically symmetric solar models, in the form of perturbations such as global and local-scale flows and structural asphericities, result in the splitting of eigenfrequencies in the observed solar spectrum. Here we describe new theoretical developments that enable the computation of sensitivity kernels for frequency splittings (a coefficients) due to general Lorentz stresses in the Sun. We draw from theoretical ideas prevalent in normal-mode coupling theory in geophysical literature to build these kernels. We plot the Lorentz-stress kernels and estimate the a-coefficients arising from a combination of deep-toroidal and surface-dipolar fields (although we note that this could equally well be substituted by another choice of Lorentz stresses). These results pave the way to formally pose an inverse problem, and infer magnetic fields from the measured even a-coefficients.