You can probably read the file stream using file listener and in processing activity save the file as .xml/.csv

Or

If you are sure about the target format, you can use parse-xml rule to read XML content and parse-delimited rule to read csv content

The 'File Listener' route is the probably the way to go : as it is can be built using standard PRPC Rules.

But if you need to go one level down: take a look at the PRPC 'Engine Docs' : there is a 'PRFile' object for instance that you use to access files on the backend server.

Yes as long as file will be accessible to server from local machine, file listener should be able to pick and process the file. Folder should have read and write permissions.

Source location

Identify a directory in which the listener looks for input files. Listener requests require read and write access to this directory.

This field supports the Global Resource Settings syntax (=PageName.PropertyName). The default property name for this field is pySourceLocation. For more information, see Using the Global Resource Settings feature.