Here's a recently discovered awesome repository on GitHub with over 130k stars, dedicated to collecting system prompts for various AI programming tools.

From Cursor to Windsurf to Trae to Devin to v0 to Lovable, it compiles all the core commands of mainstream tools you can think of into one collection.
What's the value of this? Usually, when researching prompts for a tool, you can only rely on trial and error or scattered posts.
This provides a complete map, showing how each tool defines its behavior patterns, how to enforce coding style constraints, and how to handle user instructions—all laid out in front of you, more detailed than any official documentation.
I looked through it myself and found an interesting thing: the prompt strategies of different tools vary quite a bit.
Some emphasize simplicity, others focus on process, and some even fully describe internal toolsets so you know what they can call.
This kind of information is simply not available through official channels.
x1xhlol/system-prompts-and-models-of-ai-tools
If you're developing AI programming products or looking for best practices to optimize your workflow, this repository is worth exploring thoroughly.
The power of the open-source community can really push the boundaries of commercial development.
GitHub link: