One of the more vexed issues when reconstructing ESAS completely within the SAS structure, was the handling of mask files. In practice one usually wants to detect sources over a rather broad band, in order to capture the fainter sources, but usually wants to examine the diffuse emission in some other, narrower band. Thus, the “cheese” masks created early in the process probably are not in the same energy band as the images that you are creating at this step. Thus, the names of the “cheese” masks are not going to fit easily in to the P-root-elo-ehi.fits structure, and so we didn't try to do that. What we have done is make sure that the “cheese” masks have the same coordinate frame as the other components.
One other thing we wanted avoid was requiring long lists of input file names. Thus, at this point, one will have to choose the appropriate mask files, and rename them to follow the P-root-elo-ehi.fits structure. We don't care what you use for “root”, but the prefix and the -elo-ehi.fits need to match the other component names you've created already. This is a bit of a pain, but it saves greater pain elsewhere.