what if there was a web scraping challenge that just tested a handful of randomly selected web APIs/features and checked the results you returned instead of mining bitcoin
plenty of APIs go far back to older browser, and you shouldnt (theoretically) to loop a lot, just give it random snippets that should theoretically work as one script. though im still thinking about how one would guarantee results arent made up if they don't run a browser on the server and cross-reference the results
"what about browsers/people that don't do javascript" i was thinking about this and i think i got an idea. generate an html page that links to multiple resources on seperate (sub)domains and check which ones are requested.
a competent browser is capable of inferring if, say, an element with a hidden attribute shouldnt be loaded, or even better, test how they load css. give nonexistent classes background images and see if they get fetched, write malformed css and see if referenced elements at different points across it get fetched or not, a well-done browser with a half-baked css engine will still know when and when not to load something stupid.
"what about browsers/people that don't do css either" well just fall back to the hidden/malformed elements trick, maybe try preconnect/preload links? favicon load checks? maybe iframes can be useful?
basically think randomx for web browsers but isnt wasting cpu cycles
