browser-agent · workspace

A real Chromium run on books.toscrape.com — 30 books, 13 steps, 3 categories ranked

This is the actual out/ directory the browser-agent skill produces when an agent follows the SOP and runs popularTask 2 — cross-page comparison & Top-N end-to-end. Left pane is the real file tree; the timeline reads from run.json and shows one Chromium screenshot per controller action; right pane is extracted.json rendered as a per-category list and a global Top-10 ranking. Nothing is invented — every step, every thumbnail, every price/rating cell comes from a headless Chromium that browsed the public scraping sandbox.

steps · screenshots · books extracted · 3 categories · elapsed target books.toscrape.com
navigate find_elements scroll extract done
run.json — step timeline