Generative engine optimization with Claude Code: dynamic llms.txt endpoints, AI-crawler rules in robots.txt, and JSON-LD on every page. One session.
Building 404.astro isn't enough. Without not_found_handling in wrangler.json, unknown URLs on a Workers site return a bare response instead of your page.
A static llms.txt goes stale. An Astro endpoint that reads your content collections rebuilds it every deploy. One gotcha: public/ silently shadows it.
Astro Content Collections don't hide draft posts automatically. Forgetting the filter on index and slug pages means drafts are publicly live.
TIL posts lived in their own Astro content collection. Folding them into the blog collection as a category field fixed SEO and cut duplicate logic.
