AI Briefingsystem_prompts_leaks: Original Hidden System Prompts of Major AIs Like ChatGPT and Claude
asgeirtj/system_prompts_leaks
About the project
Reveals the hidden instructions AI receives before the user's first input, exactly as they are. Provides internal rules and constraints of major services like ChatGPT, Claude, Gemini, and Grok organized into text files.
Includes prompts for coding agents like Claude Code, Codex, and Cursor, as well as various tools like Perplexity and Copilot, in addition to simple chat models. Offers a vast dataset accumulated through 771 commits, covering everything from the latest versions to older models.
Useful for researchers analyzing AI response tendencies or safety filter logic, and developers optimizing prompt engineering. Can be referenced when you want to verify specific evidence of how these systems work in actual services.
asgeirtj/system_prompts_leaks
Extracted system prompts from Anthropic - Claude Fable 5.1, Opus 5, Claude Design, Claude Code. OpenAI - ChatGPT GPT-6-Astra, Codex. Google - Gemini 3.8 Flash, 3.1 Pro, Antigravity. xAI - Grok, Grok Bot, Cursor, Kimi and more! Updated regularly.
JavaScript
This introduction was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report errors, attribution issues, or removal requests via Contact.