The X-ray Imaging Spectroscopy Mission (XRISM) provides the best spectral resolution with which to study sulfur (S) K-shell photoabsorption features from the interstellar medium (ISM). For the first time, we demonstrate the high-signal detection of interstellar atomic S II K-beta absorption in the spectrum of X-ray binaries (XRBs) 4U 1630-472 and GX 340$+$0. The persistence of this feature across multiple instruments, targets, and flux states implies that it is interstellar in nature. We measure the S II K$beta$ line centroid at $2470.8 pm 1.1$ eV after including systematic uncertainties. We also find that the most recently published high-resolution S II absorption template requires a systematic energy scale shift of $+7!-!8$ eV, which is comparable to the level of disagreement among various atomic modeling procedures. The XRISM 300 ks observation of GX 340$+$0 provides unprecedented signal-to-noise in the S K region, and we find evidence of residual absorption from solid S in the spectra of GX 340$+$0. Absorption templates from three Fe-S compounds, troilite (FeS), pyrrhotite (Fe$_7$S$_8$) and pyrite (FeS$_2$), provide equally good fits to the residuals. Even though we are not able to distinguish among these three compounds, they provide equal estimates for the abundance of S locked in dust grains. Having accounted for both the gaseous and solid S in the GX 340$+$0 sightline provides us with a direct measurement of S depletion, which is $40{%} pm 15%$. Our depletion measurement provides an upper limit to the fraction of interstellar Fe bound in Fe-S compounds of ${<}25%$, which is consistent with prior studies of Fe-S compounds via Fe L-shell absorption. Both XRBs in this study are at a distance of approximately 11 kpc and on the opposite side of the Galactic disk, suggesting that this value could represent the average S depletion of the Milky Way when integrated across all phases of the ISM.
