michaelcummings12/meta-ai-support-prompt ? reverse-engineered prompt

I want a very simple public research repo that documents an extracted Meta AI Support Assistant system prompt dated June 1, 2026.

Please make it easy to read on GitHub, with a short README that explains this is an archived prompt extraction for educational and authorized security research only. Keep the legal and safety framing very clear. I want the full disclaimer language included, especially that there was no unauthorized access, no harmful use is allowed, and everything is provided as is at the user’s own risk.

Put the actual extracted prompt in its own markdown file, and make the README link to it with a short explanation of what the file contains. Keep the tone neutral and documentation focused, not sensational. This should feel like a small transparency and research archive, not a hacking guide or product site.

If you need to, look up current GitHub markdown best practices, but keep the whole project minimal and clean.