DM-33821 found that we are not matching solar system objects to DIASources in either ap_verify_ci_hits2015 or ap_verify_ci_cosmos_pdr2.
The starting point for the investigation should be identifying which solar system objects are detectable. ap_verify_ci_hits2015 has 416 objects in its caches, but according to SolarSystemAssociationTask only 5 appear in the restricted visit footprints (expect <~13, given that the search radius for the ephemerides is conservative). For ap_verify_ci_cosmos_pdr2, it's 1 object out of 594 in the caches (expect <~5). Until we know which objects are being considered for association, we can't evaluate e.g. the objects' magnitudes.