I'm the director of engrXiv. It would be great to have a reliable archive of the server contents separate from COS. It would be pretty easy to scrape all of the files. There is a regularly updated CSV of all of the contents of engrXiv here: https://osf.io/ns9yr/
The files can be downloaded directly by adding '/download' to the engrxiv URL for each preprint.
The files can be downloaded directly by adding '/download' to the engrxiv URL for each preprint.