undefined

After installation, trigger in Claude Code with any of:

---

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

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

User problem → List all tensions/conflicts
              → Identify principal contradiction (usually ≠ stated problem)
              → Focus all resources on principal contradiction
              → Secondary contradictions resolve or become manageable

没有调查就没有发言权
No investigation, no right to speak.

Loop: Practice → Perception → Conception → Practice (verify) → repeat

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

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

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)

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

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

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

Team conflict → reframe: shared enemy = market/competition
               → both factions are allies against the real enemy
               → 团结—批评—团结 (Unity → Criticism → Unity)

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

Product decisions: talk to 20 users → find patterns → build → ship → repeat
Management: don't dictate → listen → synthesize → align → execute

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

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.

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

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.

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.

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

undefined

---

git clone https://github.com/leezythu/maoxuan-skill.git
cd maoxuan-skill

PRs welcome for: additional mental model documentation, new application patterns, corrections to framework distillation.

---