Spotted this after I fixed #1 and then manually ran the scrape.yml workflow a second time:

Cause is this default behavior:

Problem is I want wget to pick the original filename for me based on the URL, but I don't want it to avoid clobbering.