It's maybe not quite as ridiculous as it seems since a lot of these sulfide minerals are covalent and have a lot of electron sharing/delocalization.  See for example this study on carrollite where the authors assign fractional oxidation states on the basis of X-ray absorption spectra: https://www.mindat.org/reference.php?id=64588 (The sulfur is also not quite 2-, which relieves one of the cobalt atoms from having to be 4+, a fairly extreme oxidation state for cobalt)