@axichuhai: Whoa, a programming guru distilled his engineering experience into an open-source project that topped GitHub trending, with stars soaring past 90k+. The author is former Vercel engineer Mat, known for making complex tech easy to understand and involved in early Next.js development. He distilled his daily collaboration with Claud…

Holy cow, a programming guru has distilled his engineering experience into an open-source project

Topped the GitHub trending list, with stars soaring past 90k+

The author is Mat, a former Vercel engineer who contributed to early Next.js development, known for explaining complex tech in an easy-to-understand way.

He turned his daily engineering practices collaborating with Claude into 16 skills, installable with a single command, injecting decades of software engineering experience into AI.

The most acclaimed one is called Grim, which makes AI turn the tables and interrogate you—walking through a decision tree to confirm requirements one by one until it fully understands what you want before taking action.

With only 3 sentences and 42 words, many have called it “the prompt with the highest token ROI.”

Another skill specifically addresses code that won’t run: it forces the AI to first write a test that will fail, then write the minimal code to make it pass, and finally refactor. The classic red-green cycle, making it hard for AI to cut corners.

This whole project isn’t just about teaching AI how to work—it’s about establishing a continuously aligned working mechanism between you and the AI.

阿西_出海 (@axichuhai): Guys, I’ve dug up another GitHub treasure of an open-source project—hello-agents

🏆Directly topped the GitHub open-source hot list, and still climbing fast

It systematically organizes AI and Agent theory into an open-source course, covering skills like Agentic RL, SFT, GRPO—the kind you’d need when applying for high-end AI positions

Before, this content was either scattered all over or required paying for courses.