Those attachments seem not to have come through, but no matter - that's just not the supported Skyline+SpectrumMill workflow. The issue has never come up in Skyline because we just don't use those mzXML files, which I believe are generated by the Agilent tools for their internal use only.
In general regardless of what mass spec you're using you should always prefer to use the original data over any conversion format. It's more efficient and there's no risk of information loss in the translation. The mzXML format in particular is to be avoided as it is an aging standard that doesn't support recent developments like ion mobility data - all new work to support mass spec evolution is being done in mzML.
Thanks for using the Skyline support board!
Note, however, that those mzXML files do get used in building the spectrum library.
When you are using the DDA import wizard, there is a checkbox "Exclude spectrum source files", which you should check. This is the exact case this checkbox aims to support, when your spectrum source files are of a type that Skyline understands how to extract chromatograms from, but which do not really contain the necessary information to do the extraction.
In this case, those mzXML files contain only the MS/MS spectra that matched peptides in your SpectrumMill search. For chromatogram extraction, Skyline will need access to the MS1 spectra, which can only be found in your raw data files (if Agilent, then .d folders).
Once you check the checkbox, Skyline will exclude the mzXML files from its search for chromatogram extraction files and if you have .d (or other vendor) raw data files with the same basenames, then Skyline will automatically choose those for extraction. Otherwise, it will list the names and request that you point to a location that contains them.
Hope this helps. Sorry for the confusion.