This repo is a treasure trove of 20,000-plus lines for AI builders. It contains system prompts, tools, and AI models used by Claude Code, v0, Manus, Lovable, Replit, Windsurf, Warp, Devin, and more. All open source and available for study.
System Prompts and Models of AI Tools is a curated collection of real-world system prompts, tool definitions, and AI architectures from popular developer tools. It provides transparency into how production AI coding agents are built and prompted.
What You Will Find
The repository documents how major AI coding tools structure their prompts, define their tools, and organize their agent architectures. This is invaluable for anyone building their own AI agents or trying to understand the state of the art.
| Tool | Category | What Is Exposed |
|---|---|---|
| Claude Code | AI coding agent | System prompt, tool definitions |
| v0 | UI generation | Prompt structure, model config |
| Manus | General agent | Agent architecture, toolset |
| Lovable | App builder | System prompt, workflow |
| Replit | Cloud IDE | Agent integration, tools |
| Devin | Autonomous dev | Full agent architecture |
Why Study Production Prompts
Most AI development happens behind closed doors. Companies treat their system prompts as intellectual property and guard them carefully. This repository breaks that pattern by collecting and organizing prompts from tools that have chosen to open them.
Kilo Code is highlighted as a standout example: an AI coding agent with the system prompt, source code, and agent architecture all published openly. The repo encourages more tools to follow this model.
“Amazing work on this repo.” – @saboo_shubham_
“This is crazy and very insightful!” – @gargi_gupta97
What you can learn
- How production agents structure their system prompts
- Tool definition patterns that work at scale
- Error handling and recovery strategies
- Context window management techniques
- Multi-turn conversation architectures
- Prompt optimization for different model families
Project Link
If you build AI agents, this repository is a curriculum. Studying how production tools structure their prompts and architectures will teach you more than any tutorial.
If you enjoy articles about top GitHub repositories like this, don’t forget to subscribe to Technolati.com.
- How to Study System Prompts from Top AI Agents
Badge system-prompts-and-models-of-ai-tools collects 20,000+ lines of prompts, tooling, and model artifacts used across real-world agents like Claude Code, Manus, Replit,..
