Download every image a web page shows (and keep a CSV manifest)

"Grab the images from this page" sounds like a five-line script until you try it on a real article. The markup mixes img[src], srcset, lazy-load attributes, root-relative paths, and the occasional PDF