Hi,
First, we found the VIS images weird enough that we had to split them into their quadrants and solve each of those quadrants independently. Those are then quite small images, so we created the "-1" scale of index files for this purpose. I just finished building the whole set from Gaia; they are named "index-52m1-*.fits", and are available here,
For the ERO stuff, for some fields we also tried index files using different catalogs. I'm actually not sure what the minimal set would be -- we experimented with Pan-STARRS1, Legacy Surveys, and WISE. I have not yet tried to produce full-sky sets at the "-1" scale for any of these.
If you wanted to test the Gaia-based ones against a representative Euclid image set, that would be extremely helpful, since I do not have any data access rights in Euclid.
cheers,
dustin