Update on this thread:
For those who may encounter this, you may have rows where the FragmentIon column contains "precursor", "precursor [M+1]", etc, but for DIA/SRM datasets, those rows need to be removed for MSstats since MSstats performs differential abundance analysis at the fragment ion level (and not precursor level).
When exporting MSstats from Skyline, you can hit preview and filter out rows containing "precursor". See screenshots.
Tony