Replace HTML scraping with a direct call to bff.zipair.net/v1/information
via FlareSolverr. The BFF returns clean JSON (id, title, category,
publishedAt) — no regex parsing of rendered pages. Also adds first-run
seeding to avoid alerting on all historical notifications.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
