This repository collects and archives "leaked" system prompts from various AI models and platforms, including popular LLMs like Anthropic's Claude, OpenAI's ChatGPT, Google Gemini, and others. It serves as a resource for researchers, developers, and users interested in understanding the underlying instructions that guide AI behavior and output.

How It Works

The project functions as a curated repository of text files, each containing a specific system prompt. These prompts are presented as-is, without modification or interpretation. The value lies in the aggregation of these prompts, allowing for comparative analysis of how different AI providers configure their models' foundational instructions.

Quick Start & Requirements

• Install/Run: No installation required. Access prompts by cloning the repository or downloading individual files.
• Prerequisites: Basic familiarity with AI model prompting and text file handling.
• Links: GitHub Repository

Highlighted Details

• Extensive collection covering numerous AI models and versions.
• Prompts are dated, allowing for tracking of prompt evolution.
• Includes prompts for both general chat and specialized AI agents (e.g., coding assistants, IDE agents).

Maintenance & Community

The repository appears to be community-driven, with contributions likely coming from individuals discovering and sharing these prompts. There are no explicit mentions of maintainers, community channels, or a roadmap.

Licensing & Compatibility

The repository does not specify a license. The nature of "leaked" content may imply potential legal or ethical considerations regarding its distribution and use. Users should exercise caution and verify usage rights.

Limitations & Caveats

The prompts are presented without verification of their authenticity or current applicability, as AI providers frequently update their systems. The "leaked" nature of the content may also carry legal risks depending on the source and jurisdiction.