[jehrr]

Leftovers from etsy-scraper (the donor this repo was bootstrapped from) that describe Etsy as if it were this site.

Fixed

• .github/ISSUE_TEMPLATE/bug_report.yml — the pre-filing tips described Etsy's DataDome (403 to datacentre IPs, t=bv), shop_rating, and Etsy's /fr/ → /de/ redirects; the example command used an etsy.com search URL and --out grills; the "expected" placeholder was Farfetch's "96 products". Replaced with what this README/TROUBLESHOOTING measure: a scored address gets no response at all (no challenge, so a key does not help; any residential exit clears it), category listings print no rating/sold/was-price, image_url/slug_id partial by design; example uses the README's kopi-bubuk category URL and the README's 119-row healthy run.
• site_changed.yml — "Etsy changed something", etsy.com, "JSON-LD first", a lowest_price_30d tile-overlay option (no such column here), a "below 90%" DOM confirmation option and an etsy.com placeholder. Now describes the DOM-only anchors (grid container, tile marker, reading order; window.__cache on product pages) and the per-page-kind confirmation floor from CONTRIBUTING; evidence asks for --dump-html.
• SECURITY.md — "Etsy changing its markup is expected" → Tokopedia.
• CONTRIBUTING.md — removed the "rating is the LISTING's and shop_rating is the SELLER's" property: this repo has no shop_rating column and the 825/375/16,679 figures are Etsy's. Dropped the stale "Five properties … all five" count (the list has eight). "the registration form" → "a registration or login form".
• CONTRIBUTING.md (second commit) — the site-change section described a structured-price confirmation share with Etsy's per-page-kind floors (search 8%, category 70%, shop 80%). Tokopedia listings carry no structured price and the engines deliberately log plain price coverage against PRICE_FLOOR (95%); the paragraph and the template option now say that.
• captcha_solver.py — docstring pointed at a "No DataDome solver" section that went with the inherited etsy-scraper solver in 0.1.0; now says so and points at the CHANGELOG.
• smoke_test.py — privacy-pattern comment said "MediaMarkt's pages embed…" as if this site's; now names mediamarkt-scraper as a sibling.
• output_writer.py — a comment used shop_rating as its example sidecar extra; it now names scroll, which this repo's engines actually write.

Kept, deliberately

• README.md:34 — "www.etsy.com 403, www.mediamarkt.de 403 … from the same machine in the same minute" is a measured contrast naming other sites as other sites.
• Sibling provenance comments (MediaMarkt's instalment line (§4), mediamarkt-scraper's first live run, amazon-scraper's situation, the farfetch note in captcha_solver.py, the craigslist .gitignore note) — they name the sibling and state the lesson.
• Generic "a DataDome shell, a branded 403" lists in page_flow.py/product_parser.py/playwright_scraper.py — they enumerate kinds of refusal, not Etsy's behaviour.
• README badges were already complete and in family order, with an honest needs a residential exit; repo description/topics have no donor words. Released CHANGELOG sections untouched.

Verification

python3 smoke_test.py exit 0 before and after (same two engine groups skipped locally); ci_checks.py --secret-check and --help-check --sample-check exit 0; issue templates parse as YAML.

🤖 Generated with Claude Code