The Lagrangian contains a 'contact potential' , which Fermi assumed is a contact coupling of two vector currents. (prior to when parity breaking was discovered. This is now part of the Standard Model, in some cryptic form.
IMHO, you can not see it obviously in the Lagrangian until you specify the Dirac Spinors explicitly and evaluate the invariant matrix elements The Weak Interaction Hamiltonian element is
\mathcal{H}(x)=-\dfrac{G_{F}}{\sqrt{2}}\left[J^{\mu}(x)L^{+}_{\mu}(x)+h.c.\right]where
J^{\mu} and L_{\mu}
are (in modern parlance) the Hadron and Lepton currents, resp.,
The question is really how to evaluate the invariant matrix element -- at the surface of the nucleus, or at the origin ? That is, say, given some expression that appears in a cross section or a rate
\sum_{fi}\big\vert\mathcal{M}_{fi}\left[p_{ep}\rightarrow i\nabla)\right]\psi_{ep}(\mathbf{x})\big\vert_{\mathbf{x}=0}\big\vert^{2}
or would we compute the electron density at the surface ?
for example, we do compute the rate of Electron Capture , a typical weak process, we can write the capture rate as
\Gamma_{EC}=\big\vert\Psi_{ep}(0)\big\vert^{2}\mathbf{v}^{in}_{ep}\sigma_{EC}
where \big\vert\Psi_{ep}(0)\big\vert^{2} is the electron density at the origin, representing the nuclear charge.
Or , again, should this be the charge on the surface of the nucleus ?