humanlayer/12-factor-agents — reverse-engineered prompt

Reverse engineered prompt

Build me a polished open source project called 12 Factor Agents, basically a practical guide for people who want to build reliable AI agent products that real customers can use.

I want a clean landing page that explains the idea in plain English, plus a visual grid or card based nav for all 12 factors. Each factor should have its own readable page with examples and a strong opinionated point of view. The themes should cover things like turning natural language into tool calls, owning prompts and context, treating tools like structured outputs, managing state well, pause and resume flows, human handoffs, control flow, error handling, small focused agents, triggering from different places, and making agents behave more like predictable software.

It should feel more like a thoughtful handbook than a hypey AI demo. Add supporting images, workshop or example content where it makes sense, and make the repo easy to browse and extend. If you need details, look up the current public 12 Factor Agents project and docs online, then recreate something faithful and polished.

Want more depth? Deep Reverse