Scrape and analyze the user's YouTube watch history into a personal stats report focused on developer-themed watching behaviour. Use when the user asks to analyze their own YouTube **watch history** / viewing stats — e.g. "analyze my youtube watch history", "how much developer/technical content do I watch", "am I watching too many shorts", "what kind of youtube viewer am I", "break down my youtube history by topic / channel / video style", or "my youtube watch-time stats". (For the user's *own* history — it uses their signed-in YouTube account.) Sets up Playwright's agent CLI as needed and reads the last ~30 days by driving a real browser with a dedicated profile the user signs into once (it never touches their main browser profile), then writes one stats-report.md covering the developer/technical breakdown (sub-topics and styles), other topics, top channels, and watch-time estimates. Trigger even when the user doesn't say "Playwright" or "scrape" — they just want their YouTube stats.
2026-06-26