asgeirtj/system_prompts_leaks ? reverse-engineered prompt
Reverse engineered prompt
I want to build a simple website that acts as a public archive for all the "leaked" system prompts from popular AI models. The main page should be a clean, easy-to-read directory. Let's organize it by the company that makes the AI—like OpenAI, Google, Anthropic, and xAI—and then list out their specific models, such as GPT-5.4, Claude Opus, or Gemini Pro.
It should use simple tables to list each model, with a link that takes you to a page showing the full text of its system prompt. No database or complex backend is needed; it's really just a collection of text files presented in a clean way. It should be easy for me to add new prompts as they are discovered.
Are you gonna build this?
make sure you review the code using coderabbit