Your workflow is incorrect for what you ask.
I'm assuming that the URLs in the source page are the .PDFs you're trying to download, so that's OK.
The issue is that the New Text File action returns the file that was created to contain the files.
This is passed into Extract URLs from Text. You are expecting this action to see the text written to the file. It does not. It sees the path of the file that was created.
You have two options - you can pass the output from Filter URLs directly into the Download URLs action (i.e. there is no need to write to the intermediate file.
If you do need the intermediate file (i.e. a log of files downloaded), then you need an additional action to read the file back in and pass the contents of the file to the Download URLs action.