The Five Core Mental Models in Zhangxuefeng-Skill Explained

The Zhangxuefeng-Skill encodes five cognitive frameworks—Social Filter Theory, Choice over Effort, Employment Reverse-Engineering, Class Realism, and Controversy Drives Propagation—that serve as the decision-making engine for career and education advice.

The zhangxuefeng-skill repository (alchaincyf/zhangxuefeng-skill) provides an AI skill that applies a unique cognitive framework to career counseling. According to the source code in SKILL.md, the five core mental models form the "operating system" that drives every response, filtering raw data through a pragmatic lens designed for real-world outcomes.

The Five Core Mental Models Defined

The cognitive framework is explicitly defined in SKILL.md, where each model occupies a specific line range and serves a distinct function in the reasoning pipeline.

1. Social Filter Theory (社会筛子论)

Defined in SKILL.md at lines 30-33, Social Filter Theory posits that society operates as a giant sieve, filtering individuals by concrete markers such as education credentials, property ownership, and job titles. The skill uses this model to immediately assess a user's starting position within social hierarchies before recommending career paths.

2. Choice Over Effort (选择 > 努力)

Located in SKILL.md at lines 43-46, this model asserts that the right direction matters far more than the volume of effort. A bad choice wastes effort regardless of intensity. The skill applies this heuristic to prevent users from selecting high-effort, low-return career tracks based on prestige alone.

3. Employment Reverse-Engineering (就业倒推法)

Found in SKILL.md at lines 56-60, Employment Reverse-Engineering mandates selecting fields by looking backward from real employment data—specifically median outcomes and placement rates—rather than forward from academic prestige or personal interest. This model forces data-driven field selection based on economic survivability.

4. Class Realism (阶层现实主义)

Defined in SKILL.md at lines 69-73, Class Realism establishes a hierarchy of priorities: if you do not possess generational wealth ("mines"), you must first secure a livelihood before chasing ideals. The skill uses this as a veto mechanism against high-risk aspirational advice for economically vulnerable users.

5. Controversy Drives Propagation (争议即传播)

Located in SKILL.md at lines 82-86, this model acknowledges that mediocre advice is forgotten while extreme, debatable viewpoints generate maximum reach. The skill occasionally deploys this framework to frame advice in deliberately provocative terms that cut through information noise, ensuring the core message achieves visibility.

How the Mental Models Power the Decision Engine

These five models do not operate in isolation; they constitute the engine that shapes the skill’s three-step decision-making flow:

1. Problem classification – The skill first determines whether a question requires fresh data or can be answered purely through the mental models.
2. Data-driven research – When concrete facts are required, the skill fetches up-to-date employment or industry data (Step 2 in the workflow).
3. Model-driven reasoning – Once facts are gathered, or for pure-framework questions, the skill applies the five mental models to produce a decisive, unambiguous answer (Step 3).

The README.md file (lines 28-37) reinforces this architecture by presenting a summary table of the five models, confirming their central role in the skill's design.

Implementation and Usage Examples

The skill is designed for any Agent-Skills-compatible runtime (Claude Code, Codex, Cursor, etc.). Below are minimal snippets demonstrating how the five mental models activate in practice.

Direct Invocation via CLI

# Install the skill (once)

npx skills add alchaincyf/zhangxuefeng-skill

# Ask a question that activates the mental-model engine

echo "我孩子想学金融，适合吗？" | npx skills run zhangxuefeng-perspective

Typical output (truncated):

我跟你说，先问下家里情况。你家有没有矿？…
采用 **社会筛子论** → 金融行业对普通家庭风险大。…
用 **就业倒推法** 看 2025 年金融专业毕业生中位数薪资…  
结论：别碰金融，选计算机或电气更靠谱。

The response interleaves the five mental models (社会筛子论, 就业倒推法, 阶层现实主义, 选择>努力, and occasionally 争议即传播 for emphasis).

Using the Skill via HTTP (OpenAI-Compatible API)

{
  "model": "alchaincyf/zhangxuefeng-skill",
  "messages": [
    {"role": "user", "content": "我想转行做AI工程师，值得吗？"}
  ]
}

Typical reply:

我跟你说，先查下 AI 工程师的 2026 年就业率和中位数薪资……
- **就业倒推法**：普通毕业生五年后平均年薪 15 万，增长趋缓。  
- **选择 > 努力**：如果你家没矿，别盲目冲，先把基础功打好。  
- **争议即传播**：目前行业争议大，切记别被热点冲昏头。
结论：AI 有潜力，但对普通家庭而言风险仍高，建议先选计算机+AI 组合。

Both examples showcase how the skill automatically layers the five mental models onto the answer, converting raw data into the skill's characteristic pragmatic advice.

Key Source Files

The five mental models and their implementation are distributed across these critical files:

• SKILL.md – Core definition of the five models (lines 30-86), workflow logic, role-play rules, and decision heuristics.
• README.md – High-level overview containing the "5个心智模型" summary table (lines 28-37).
• examples/demo-conversation.md – Real-world conversation logs demonstrating the models in action.
• references/research/05-decisions.md – Source research that fed the mental-model extraction.

These files together constitute the knowledge base that powers the five core mental models and the overall behavior of the Zhangxuefeng-Skill.

Summary

• Social Filter Theory (SKILL.md lines 30-33) evaluates social position through education, housing, and job filters.
• Choice Over Effort (SKILL.md lines 43-46) prioritizes directional correctness over raw effort volume.
• Employment Reverse-Engineering (SKILL.md lines 56-60) selects fields based on median employment outcomes, not prestige.
• Class Realism (SKILL.md lines 69-73) mandates securing livelihood before pursuing ideals for those without wealth.
• Controversy Drives Propagation (SKILL.md lines 82-86) leverages extreme viewpoints to maximize message reach and retention.

Frequently Asked Questions

What is the primary purpose of the five core mental models in Zhangxuefeng-skill?

The five core mental models serve as the skill's "operating system," converting raw career data into pragmatic, class-aware advice. They act as decision filters that prioritize economic survival and social mobility over aspirational or prestige-based recommendations.

How does Employment Reverse-Engineering differ from traditional career advice?

Traditional advice often proceeds forward from interest or prestige, while Employment Reverse-Engineering (defined in SKILL.md lines 56-60) works backward from verified median salary and employment outcome data. This ensures recommendations are grounded in economic reality rather than academic reputation.

Where are the five mental models defined in the source code?

All five models are explicitly defined in SKILL.md within specific line ranges: lines 30-33 (Social Filter), lines 43-46 (Choice Over Effort), lines 56-60 (Employment Reverse-Engineering), lines 69-73 (Class Realism), and lines 82-86 (Controversy Drives Propagation). The README.md (lines 28-37) provides a summary table for quick reference.

Can I modify or extend these mental models for my own use case?

Yes. Since the skill is open-source, you can fork the repository and edit SKILL.md to adjust the model definitions or add new cognitive frameworks. However, modifying the line references will require updating any dependent documentation that cites specific model locations.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}