ed-donner/agents — reverse-engineered prompt

Reverse engineered prompt

GitHub

Build me a complete learning workspace for a 6 week course on autonomous AI agents. I want it to feel like a friendly course repo that someone can open in Cursor and work through step by step, mostly with Jupyter notebooks.

Organize it by weeks, starting with foundations, then OpenAI agent examples, CrewAI, LangGraph, AutoGen, and MCP. Include practical exercises and small demo agents people can run, not just explanations. Add clear setup instructions for Windows, Mac, and Linux, including how to handle API keys, Python setup, and common install problems. Also include helpful guide notebooks, troubleshooting notes, and reminders about API costs, with options for cheaper or local models like Ollama.

Keep the tone encouraging and beginner friendly, like a course instructor wrote it. If you need current details for any agent framework, look up the latest docs online and make the examples runnable.

Want more depth? Deep Reverse