HTTP 403 from curl != broken site

When probing dozens of URLs with curl --max-time, a status of 403 usually means bot protection (Cloudflare, AWS WAF, ModSec), not a dead page — those URLs work fine in a real browser. Always set a real-browser User-Agent header before drawing conclusions; raw curl exit codes alone overcount failures. Status 000 = DNS/connect failure (genuinely broken or wrong domain). Status 404 = wrong path on a working domain (try the homepage or other paths). Only 000 and 4xx-besides-403 are actually-broken signals.