SKILL.md
$27
After installation, trigger in Claude Code with any of:
毛选
教员
用毛选的方式分析
从毛选的角度
教员怎么看
用毛选框架
What the Skill Does
The skill loads SKILL.md into your AI agent's context, giving it:
- 7 Mental Models — operational frameworks extracted from core writings
- 10 Decision Heuristics — concrete rules for ambiguous situations
- Expression DNA — the rhetorical style (direct address, historical analogy, dialectical structure)
- Honest Boundaries — what it can and cannot do
The agent does NOT roleplay as Mao. It uses his analytical frameworks to examine your problem.
Repository Structure
maoxuan-skill/
├── SKILL.md # Core skill file loaded by agent
├── README.md
├── LICENSE # MIT
└── references/
└── research/
├── 01-core-writings.md # Distilled core texts
├── 02-strategic-thinking.md # Strategic thinking analysis
├── 03-expression-dna.md # Expression style DNA
├── 04-external-views.md # External perspectives
├── 05-decisions.md # Key decision analysis
└── 06-timeline.md # Timeline reference
The 7 Core Mental Models
1. 矛盾分析法 — Contradiction Analysis
Source: 《矛盾论》 (On Contradiction)
Find the principal contradiction among all contradictions. Resolving it moves everything else.
Problem has N issues → identify the ONE that, if resolved, unlocks the others
Don't fight symptoms → fight the root contradiction
Principal aspect vs. secondary aspect within each contradiction
Application pattern:
User problem → List all tensions/conflicts
→ Identify principal contradiction (usually ≠ stated problem)
→ Focus all resources on principal contradiction
→ Secondary contradictions resolve or become manageable
2. 实践认识循环 — Practice-Cognition Loop
Source: 《实践论》 (On Practice)
没有调查就没有发言权
No investigation, no right to speak.
Loop: Practice → Perception → Conception → Practice (verify) → repeat
Application pattern:
Don't theorize without data → go investigate first
Hypothesis → smallest possible test → update model → next test
Never skip the "go talk to real users/customers" step
3. 持久战略 — Protracted War Strategy
Source: 《论持久战》 (On Protracted War)
Three phases when you're the weaker party:
Phase 1: Strategic defensive — preserve strength, avoid decisive battles you'll lose
Phase 2: Strategic stalemate — attrit the enemy, build your base
Phase 3: Strategic offensive — counterattack from position of strength
Application pattern:
Early startup: don't fight big corp head-on (Phase 1)
Growing startup: find your niche, deepen it, build defensible moat (Phase 2)
Established player: expand from stronghold (Phase 3)
4. 农村包围城市 — Rural Encirclement of Cities
Source: 《星星之火,可以燎原》 (A Single Spark Can Start a Prairie Fire)
Don't attack the enemy's strongest point
Find underserved edges the dominant player ignores
Build an unassailable base there
Expand concentrically until you can challenge the center
Application pattern:
Big corp's weakness: can't focus deeply on any single niche
Your advantage: you can go 10x deeper in ONE narrow segment
Win that segment completely → expand to adjacent segment → repeat
5. 统一战线 — United Front
Source: 《中国社会各阶级的分析》
把朋友搞得多多的,把敌人搞得少少的
Maximize allies, minimize enemies.
Distinguish: Who is the real enemy? Who is a potential ally even if uncomfortable?
Don't make enemies unnecessarily
Isolate the primary enemy by uniting everyone else
Application pattern:
Team conflict → reframe: shared enemy = market/competition
→ both factions are allies against the real enemy
→ 团结—批评—团结 (Unity → Criticism → Unity)
6. 群众路线 — Mass Line
Source: 《关于领导方法的若干问题》
从群众中来,到群众中去
From the masses, to the masses.
Gather scattered ideas from users/team
Synthesize into coherent direction
Return to users/team as clear guidance
Verify in practice → iterate
Application pattern:
Product decisions: talk to 20 users → find patterns → build → ship → repeat
Management: don't dictate → listen → synthesize → align → execute
7. 纸老虎论 — Paper Tiger Theory
战略上藐视敌人,战术上重视敌人
Strategically: the enemy is a paper tiger (don't be paralyzed by their size)
Tactically: take them seriously in every specific engagement
Big corp LOOKS invincible → actually has massive structural weaknesses
You LOOK tiny → actually have massive agility advantages
10 Decision Heuristics
Heuristic
When to Apply
没有调查就没有发言权
Before any major decision — go gather data first
抓主要矛盾
When overwhelmed by multiple problems
星星之火,可以燎原
When tempted to scale before the core is solid
一分为二看问题
When analysis seems one-sided
不打无准备之仗
Before launching any initiative
自力更生为主,争取外援为辅
Resource allocation decisions
集中优势兵力,各个歼灭
When to concentrate vs. spread resources
在战略上藐视,在战术上重视
Facing intimidating competition
从群众中来,到群众中去
Product/org decisions lacking user grounding
把朋友搞得多多的,把敌人搞得少少的
Stakeholder and coalition decisions
Usage Patterns
Pattern 1: Competitive Strategy Analysis
Trigger: 毛选
Problem: We're a 10-person startup. Google just shipped a product
that does what we do.
Framework applied:
→ 纸老虎论: Google's size = paper tiger. They have 1000 priorities.
You have 1.
→ 农村包围城市: Don't fight Google's product. Find the vertical
they'll never care about. Go 10x deeper there.
→ 持久战略 Phase 1: Don't go head-to-head. Survive, build base.
→ 矛盾分析: Principal contradiction = you haven't found your
unassailable niche yet. That's the ONE thing to solve.
Pattern 2: Team Conflict Resolution
Trigger: 教员怎么看
Problem: Two tech leads hate each other. Team is split.
Shipping has stopped.
Framework applied:
→ 矛盾分析: Principal contradiction = missed deadlines/market window.
Their conflict = secondary contradiction.
→ 统一战线: They are NOT enemies of each other. Market is the enemy.
Unite them against the real enemy.
→ 团结—批评—团结:
Step 1: Establish shared goal (the "抗日" = shipping Q3)
Step 2: Honest criticism session — facts on the table
Step 3: Return to unity around shared goal
Pattern 3: Career/Pivot Decisions
Trigger: 用毛选的方式分析
Problem: Backend engineer, 5 years experience,
wants to enter AI but feels outgunned.
Framework applied:
→ 纸老虎论: "AI market is saturated" = paper tiger.
Most people are doing surface-level work.
→ 根据地思维: Your backend experience IS your base.
AI engineers don't have it.
→ 实践认识循环: Before strategizing — investigate.
What are the actual bottlenecks to AI adoption
in your industry? (hint: usually not the model)
→ 农村包围城市: Don't compete on algorithms.
Own the "AI + [your industry]" niche.
Pattern 4: Resource Allocation
Trigger: 毛选
Problem: 3 possible product directions, resources for 1.
Framework applied:
→ 矛盾分析: Which direction resolves the PRINCIPAL contradiction
(survival/growth)?
→ 集中优势兵力: Don't spread. Concentrate everything on ONE.
→ 持久战略: Which direction gives you a defensible base
to expand from later?
→ 没有调查就没有发言权: Have you actually talked to users
in each direction? Do that first.
Expression Style (What the Skill Sounds Like)
The framework uses a distinctive rhetorical structure:
1. Direct address: "同志" / "你这个问题..."
2. Historical grounding: brief concrete analogy (not lecture)
3. Problem reframe: "你以为主要矛盾是X,其实是Y"
4. Framework application: named model + your specific situation
5. Concrete next action: specific, not vague
6. Closing formulation: aphorism tied to the analysis
It does NOT:
- Lecture on history at length
- Repeat quotes without applying them
- Give generic motivational content
- Avoid naming the hard thing
Skill Compatibility
Agent
Compatible
Claude Code
✅ Primary target
Cursor
✅ via SKILL.md in context
Codex
✅
skills.sh
✅
# Verify skill is loaded
npx skills list
# Should show: maoxuan-skill-cognitive-framework
Honest Boundaries
What the skill can do:
- Apply contradiction analysis to decompose complex problems
- Use protracted war / rural encirclement for competitive strategy
- Apply mass line thinking to product and org decisions
- Use united front logic for stakeholder/coalition mapping
- Provide dialectical (一分为二) analysis that avoids one-sidedness
What it cannot do:
Limitation
Reason
Replace actual domain expertise
Framework ≠ knowledge of your specific market
Generate specific execution plans
Provides analytical structure, not implementation
Poetry / literary creation
Style is extractable, literary genius is not
Historical evaluation
Skill focuses on methodology, not historical judgment
Guarantee outcomes
Analysis improves decisions, doesn't determine results
Why This Framework for Technical/Startup Problems
Most business frameworks assume you have resources. This one was built for the opposite:
Framework
Starting conditions
Jobs/Apple
Talent + capital + distribution
Musk/Tesla
Capital + regulatory navigation
Munger/Berkshire
Information asymmetry + long time horizon
毛选
Extreme resource disadvantage vs. vastly stronger opponent
If you're a small team facing a large competitor, a junior person navigating a large org, or a founder with limited runway — this framework was literally built for your situation.
Contributing
git clone https://github.com/leezythu/maoxuan-skill.git
cd maoxuan-skill
# Core skill file
vim SKILL.md
# Research references
vim references/research/01-core-writings.md
PRs welcome for: additional mental model documentation, new application patterns, corrections to framework distillation.
License
MIT — use freely, modify freely, build on it freely.
语录告诉你他说过什么。毛选.skill 帮你用他的方式看你的问题。
星星之火,可以燎原。