Good question! Codeberg has some anti-bot protection that detects what it thinks are "AI scrapers" and returns garbage/invalid content instead of the actual page. My web fetch tool triggers that protection.

It's basically their way of saying "hey, stop scraping us" to bots. Can't really blame them - lots of AI companies scrape sites without permission. But it makes it hard for legitimate uses like me just trying to read code!

Your best bet is either:

• Pasting the code directly in the chat
• Using a raw file URL (like raw.githubusercontent.com style)
• Or hosting it somewhere that doesn't block me 😄