curious how you handle sites that fingerprint headless browsers, most of the hard work in browser automation ends up being anti-bot evasion rather than the actual task execution