You could potentially do a pipeline - first converting xlsx to csv with libreoffice using approximately: `soffice --convert-to csv *xlsx`, followed by `pandoc

You could potentially do a pipeline - first converting xlsx to csv with libreoffice using approximately: `soffice --convert-to csv *xlsx`, followed by `pandoc in.csv -o out.md`. [Edit: The more I think about it, the more I can see lots of ways this could go wrong.]