My sense is you know that the neutron is a composite particle of size 0.8 fm, and wonder how a tree photon interaction by anything charged inside the "upper" n (it's actually simpler to think of e-n scattering) can "see" the quarks of the lower n, but fail to "see" the whole as a neutral object.
The point is that for energies much lower than 1/(0.8fm) you see the total, aggregate charge of the n, null, because you must add the lowest-order tree diagrams you wrote, in this case the photon coupling to the u, plus to the d, plus to the other d valence quark. The resulting coupling, if you cannot resolve the difference among these three valence quarks, is just (2/3-1/3-1/3)=0! (This is something like the chiral bag model.)
The comments and the other answer remind you that if, at slightly higher energies, comparable to the inverse Compton wavelength of the n, you can resolve the three valence quarks, (they have a peculiar spin structure, where they contribute unequally), you can "see" their magnetic effects and the neutral n has a nontrivial magnetic moment, about -2/3 that of the charged proton! (This is because magnetism is shorter-ranged than electric forces.)
In fact, as you increase your scattering energy to come close to the above-mentioned radius of the n, you start seeing inhomogeneities in the charge distribution inside this neutral particle! The charge-radius-squared of the neutron is -0.1 fm², which, of course, means the negatively-charged d quarks are radially distributed "further out" in the neutron compared to the u quark", so to speak. The corresponding quantity for the proton is 0.6 fm². So for such higher energies, you indeed have a mismatch between the hadron and quark pictures: that's why people bother with the latter...