gitmyhub

zhangxuefeng-skill

★ 8.5k updated 24d ago

张雪峰.skill — 张雪峰的认知操作系统。高考志愿/考研/职业规划的实战思维框架。由女娲.skill生成。

This repository is a Claude Code skill that lets you interact with an AI modeled on the reasoning style of Zhang Xuefeng, a well-known Chinese education commentator famous for giving blunt, statistics-driven advice about college major selection and career planning. Rather than just mimicking his speech, the skill tries to reproduce his underlying decision frameworks so the AI can apply them to new questions you bring.

Zhang Xuefeng is known in China for views like "society is a giant sieve" (filtering people by degree, housing, and job status), prioritizing choosing the right major over choosing a prestigious school, and working backward from job market data to evaluate whether a field is worth entering. The skill encodes five of his core mental models, eight specific decision heuristics he uses repeatedly, and detailed notes on his speaking style, including his short punchy sentences, blunt assessments, and northeastern Chinese dialect flavor.

The README shows example conversations: a parent asks whether their child with a particular college entrance exam score should study finance, and the AI responds the way Zhang Xuefeng would, asking about family background in the industry, citing median employment data for finance graduates, and redirecting toward computer science or electrical engineering instead. Another example handles whether to pursue a graduate degree from a top school, and another addresses whether to study art when your family has no money.

To use it, you install the skill with one command via npx and then invoke it inside Claude Code by asking questions framed around Zhang Xuefeng's perspective. The skill itself is a SKILL.md file accompanied by six research documents covering his published books, recorded interviews, media profiles, and a documented timeline of his career decisions.

The skill was generated automatically by a related tool called Nuwa.skill, which takes a person's name, runs multiple research agents in parallel, distills a cognitive model, and assembles a deployable skill file. Several similar persona skills exist in the same author's GitHub account, covering Steve Jobs, Elon Musk, Charlie Munger, Richard Feynman, and others.