Kutuyyy / Leaked-System-Prompt-AI

# 🔍 AI Systems Transparency & Observability
> **Making AI systems more transparent, understandable, and observable.**
---
## 📖 About
This project is dedicated to **AI systems transparency, research, and observability**.
It collects and documents system prompts, behavioral instructions, tool definitions, guidelines, and other system-level information from major AI models and agent platforms.
The goal is simple:
> **Understand how AI systems are instructed, configured, and guided — not just what they output.**
---
> ⚠️ **DO THIS AT YOUR OWN RISK.**
>
> Make sure you understand what you're doing before using any information or tools from this repository with your AI account.
---
## 🌐 Systems & Platforms
This project covers research and documentation related to major AI systems and agent platforms, including:
| Platform | Category |
|:---|:---|
| 🤖 **OpenAI** | AI Models & Assistants |
| 🔵 **Google** | AI Models & Agents |
| 🟣 **Anthropic** | AI Models |
| ⚫ **xAI** | AI Models |
| 🟠 **Perplexity** | AI Search & Agents |
| 🟢 **Cursor** | AI Coding Agent |
| 🌊 **Windsurf** | AI Coding Agent |
| 🧠 **Devin** | AI Software Engineer |
| 🟡 **Manus** | AI Agent |
| 🟦 **Replit** | AI Development Platform |
| ➕ **Others** | AI Models & Agents |
> The list is continuously evolving as new systems and research become available.
---
# 🎯 Why This Exists
> **"In order to trust the output, one must understand the input."**
Modern AI systems are often controlled by multiple layers of instructions and configuration.
These layers can influence:
- 🗣️ What an AI can or cannot say
- 🎭 What persona or behavior it follows
- 🚫 How refusals and restrictions are handled
- 🔄 How the system redirects certain requests
- 🛠️ Which tools and capabilities are available
- 📋 Which behavioral guidelines are applied
- 🧠 How the system is expected to respond
- ⚖️ What safety, ethical, or policy frameworks affect its behavior
Understanding these instructions provides valuable context when studying AI behavior.
---
# 🔬 The AI Instruction Stack
A simplified representation of how an AI system may process a request:
 [code] 
The visible conversation is only one part of the overall system.
---
# 🧩 What We Document
Where available, entries may contain:
| 🔎 Information | 📝 Description |
|:---|:---|
| 🤖 **Model** | AI model or system name |
| 🏷️ **Version** | Specific model/system version |
| 📅 **Date** | Date the information was obtained |
| 🧾 **System Prompt** | Available system-level instructions |
| 📋 **Guidelines** | Behavioral and operational instructions |
| 🛠️ **Tools** | Available tools and capabilities |
| 🔧 **Configuration** | Relevant system configuration |
| 📝 **Context** | Additional observations |
| 🔗 **Sources** | Supporting references |
---
# 📂 Repository Structure
 [code] 
The structure may evolve as the project grows.
---
# 🧪 Research & Methodology
Information in this repository may come from:
- 🔍 Publicly accessible interfaces
- 🧪 Security and behavioral research
- 📚 Existing d