A GitHub repository aggregating extracted system prompts from Claude, GPT, Gemini, Grok, and other AI models—serving as a public database of AI agent instructions. It has 49.6k stars and appears to be trending via GitHub's platform, with Washington Post coverage suggesting cultural resonance around AI transparency.

One insight: This is not a productizable company or replicable wedge—it's a documentation scrape of proprietary system prompts that will trigger legal/CFAA risk for any commercial copycat, making fast-following pointless and moat-proof against competition.