Document360 articles include an AI-crawler discovery banner ("Fetch the complete documentation index at: [site]/llms.txt...") as the first visible block of body content on every article page. When an article does not have a custom SEO/meta description configured, some third-party link-unfurling tools fall back to scraping the first visible text on the page to build a link preview, and that fallback picks up this banner text instead of the article's actual content, since the banner sits ahead of the real article body in the DOM.

This produces confusing looking link previews (showing crawler instructions instead of article content) on any platform whose unfurler doesn't strictly rely on og:description/meta description tags.

Suggestion:

Exclude the llms.txt banner from any "visible body text" extraction path (e.g., mark it with aria-hidden, move it out of the main readable-content container, or otherwise flag it as non-content), so it's available to AI tooling but isn't treated as the page's lead text by generic scrapers/unfurlers.

Alternatively/additionally, auto-generate a fallback meta description from the actual article content when no manual SEO description has been set, so third-party tools have a real description to use instead of needing to scrape the body at all.