Oh those are RDL alignment marks? they looked to be aligning overglass cuts to M5

Yes, those are the ones. The first step in the WSCLP bump bonding is to align to the pads (metal5) and then etch off any oxidation residue on top, so yes, the markers are to align the topmost SkyWater metal5 mask to Micross' first overglass etch.

oh, so the wafers leave skywater fully covered in overglass and then micross does the glass etch prior to bumping?

My understanding (which could always be incorrect) is that SkyWater just prepares the wafers as usual (with their own overglass cut at the end), but that Micross makes no assumptions and will etch over the pads regardless. We had a three-way negotiation between SkyWater, Micross, and efabless as to where to put those alignment markers, and settled on the corner of the two test die areas as the best place. I generated the redistribution layout for the Caravel padframe, and then we arrayed those into a full reticle and dropped the two alignment markers at the agreed-upon spot. Looks like that all worked perfectly!