К основному содержанию
скилл CX Codex

deep-reading-analyst — скилл для Claude Code: анализ текстов

Скилл deep-reading-analyst разбирает статьи, исследования и другие длинные тексты по выбранной глубине и аналитическим моделям. Помогает проверить аргументы, сравнить источники и превратить чтение в заметки или действия.

MIT ★ 286 обновлён 2026-05-15

Установить в Claude Code · Через Skills CLI (Vercel) Все способы и платформы ↓
npx skills add YYH211/Claude-meta-skill --skill deep-reading-analyst -a claude-code
Открыть на GitHub

Что делает deep-reading-analyst

Скилл начинает с цели чтения, желаемой глубины и предпочтительных моделей. Если пользователь не отвечает, он выбирает стандартный уровень и подбирает модели по типу материала. Анализ сначала описывает структуру и главный тезис текста, затем применяет подходящие модели: от SCQA и 5W2H до критического мышления и сравнения источников.

Для работы он использует файлы из каталога references, включая материалы о SCQA, 5W2H и критическом мышлении. Стандартный уровень добавляет проверку аргументов, логических ошибок и рисков; глубокий — анализ с помощью ментальных моделей, первых принципов, системного мышления и шести шляп. В описании также предусмотрен веб-поиск на исследовательском уровне. Если для задачи важен конкретный формат разбора, его лучше назвать заранее: скилл спрашивает о предпочтительных моделях.

Какие задачи решает

  • выделить тезис и структуру длинной статьи
  • найти слабые места в аргументах и логике
  • сопоставить выводы из нескольких источников
  • составить учебные заметки и практические шаги

Примеры запросов агенту

  • Разбери эту статью и выдели её главный тезис, аргументы и пробелы.
  • Сравни эти исследования и покажи, где их выводы расходятся.
  • Проанализируй этот текст с помощью SCQA и критического мышления.

Как установить deep-reading-analyst

1Через Skills CLI (Vercel)

npx skills add YYH211/Claude-meta-skill --skill deep-reading-analyst -a claude-code

2Вручную: положить файлы

~/.claude/skills/deep-reading-analyst/SKILL.md

Папка скилла целиком; для одного проекта — .claude/skills/deep-reading-analyst/

По шагам, с командами для каждой платформы: Как установить скилл в Claude Code: пять способов

Вопросы о deep-reading-analyst

Как установить deep-reading-analyst в Claude Code?

Выполните npx skills add YYH211/Claude-meta-skill --skill deep-reading-analyst -a claude-code, затем используйте файл ~/.claude/skills/deep-reading-analyst/SKILL.md.

Что делает deep-reading-analyst?

Он структурирует статьи и исследования, проверяет аргументы и применяет модели анализа в зависимости от выбранной глубины. Можно также запросить сравнение источников или практические выводы.

Какие модели анализа использует скилл?

Для короткого разбора предусмотрены SCQA и 5W2H; стандартный уровень добавляет критическое мышление и инверсию. Глубокий уровень включает ментальные модели, первые принципы, системное мышление и шесть шляп.

Исходный текст

SKILL.md · MIT
---
name: deep-reading-analyst
description: "Comprehensive framework for deep analysis of articles, papers, and long-form content using 10+ thinking models (SCQA, 5W2H, critical thinking, inversion, mental models, first principles, systems thinking, six thinking hats). Use when users want to: (1) deeply understand complex articles/content, (2) analyze arguments and identify logical flaws, (3) extract actionable insights from reading materials, (4) create study notes or learning summaries, (5) compare multiple sources, (6) transform knowledge into practical applications, or (7) apply specific thinking frameworks. Triggered by phrases like 'analyze this article,' 'help me understand,' 'deep dive into,' 'extract insights from,' 'use [framework name],' or when users provide URLs/long-form content for analysis."
---

# Deep Reading Analyst

Transforms surface-level reading into deep learning through systematic analysis using 10+ proven thinking frameworks. Guides users from understanding to application through structured workflows.

## Framework Arsenal

### Quick Analysis (15min)
- 📋 **SCQA** - Structure thinking (Situation-Complication-Question-Answer)
- 🔍 **5W2H** - Completeness check (What, Why, Who, When, Where, How, How much)

### Standard Analysis (30min)
- 🎯 **Critical Thinking** - Argument evaluation
- 🔄 **Inversion Thinking** - Risk identification

### Deep Analysis (60min)
- 🧠 **Mental Models** - Multi-perspective analysis (physics, biology, psychology, economics)
- ⚡ **First Principles** - Essence extraction
- 🔗 **Systems Thinking** - Relationship mapping
- 🎨 **Six Thinking Hats** - Structured creativity

### Research Analysis (120min+)
- 📊 **Cross-Source Comparison** - Multi-article synthesis

## Workflow Decision Tree

```
User provides content
    ↓
Ask: Purpose + Depth Level + Preferred Frameworks
    ↓
┌─────────────────┬─────────────────┬─────────────────┬─────────────────┐
│   Level 1       │   Level 2       │   Level 3       │   Level 4       │
│   Quick         │   Standard      │   Deep          │   Research      │
│   15min         │   30min         │   60min         │   120min+       │
├─────────────────┼─────────────────┼─────────────────┼─────────────────┤
│ • SCQA          │ Level 1 +       │ Level 2 +       │ Level 3 +       │
│ • 5W2H          │ • Critical      │ • Mental Models │ • Cross-source  │
│ • Structure     │ • Inversion     │ • First Princ.  │ • Web search    │
│                 │                 │ • Systems       │ • Synthesis     │
│                 │                 │ • Six Hats      │                 │
└─────────────────┴─────────────────┴─────────────────┴─────────────────┘
```

## Step 1: Initialize Analysis

**Ask User (conversationally):**
1. "What's your main goal for reading this?"
   - Problem-solving / Learning / Writing / Decision-making / Curiosity
2. "How deep do you want to go?"
   - Quick (15min) / Standard (30min) / Deep (60min) / Research (120min+)
3. "Any specific frameworks you'd like to use?"
   - Suggest based on content type (see Framework Selection Guide below)

**Default if no response:** Level 2 (Standard mode) with auto-selected frameworks

### Framework Selection Guide

Based on content type, auto-suggest:

```markdown
📄 Strategy/Business articles → SCQA + Mental Models + Inversion
📊 Research papers → 5W2H + Critical Thinking + Systems Thinking
💡 How-to guides → SCQA + 5W2H + First Principles
🎯 Opinion pieces → Critical Thinking + Inversion + Six Hats
📈 Case studies → SCQA + Mental Models + Systems Thinking
```

## Step 2: Structural Understanding

**Always start here regardless of depth level.**

### Phase 2A: Basic Structure

```markdown
📄 Content Type: [Article/Paper/Report/Guide]
⏱️ Estimated reading time: [X minutes]
🎯 Core Thesis: [One sentence]

Structure Overview:
├─ Main Argument 1
│   ├─ Supporting point 1.1
│   └─ Supporting point 1.2
├─ Main Argument 2
└─ Main Argument 3

Key Concepts: [3-5 terms with brief definitions]
```

### Phase 2B: SCQA Analysis (Quick Framework)

Load `references/scqa_framework.md` and apply:

```markdown
## SCQA Structure

**S (Situation)**: [Background/context the article establishes]
**C (Complication)**: [Problem/challenge identified]
**Q (Question)**: [Core question being addressed]
**A (Answer)**: [Main solution/conclusion]

📊 Structure Quality:
- Clarity: [★★★★☆]
- Logic flow: [★★★★★]
- Completeness: [★★★☆☆]
```

### Phase 2C: 5W2H Completeness Check (if Level 1+)

Quick scan using `references/5w2h_analysis.md`:

```markdown
## Information Completeness

✅ Well-covered: [What, Why, How]
⚠️  Partially covered: [Who, When]
❌ Missing: [Where, How much]

🔴 Critical gaps: [List 1-2 most important missing pieces]
```

## Step 3: Apply Thinking Models

**Select based on depth level and user preference:**

### Level 1 (Quick - 15 min)
**Core**: Structure + SCQA + 5W2H Quick Check

Output:
- SCQA breakdown
- Information gaps (from 5W2H)
- TOP 3 insights
- 1 immediate action item

### Level 2 (Standard - 30 min)
**Add**: Critical Thinking + Inversion

Load and apply:
- `references/critical_thinking.md`:
  - Argument quality assessment
  - Logic flaw identification
  - Evidence evaluation
  - Alternative perspectives

- `references/inversion_thinking.md`:
  - How to ensure failure? (reverse the advice)
  - What assumptions if wrong?
  - Missing risks
  - Pre-mortem analysis

```markdown
## Critical Analysis

### Argument Strength: [X/10]
Strengths:
- [Point 1]

Weaknesses:
- [Point 1]

Logical fallacies detected:
- [If any]

## Inversion Analysis

🚨 How this could fail:
1. [Failure mode 1] → Mitigation: [...]
2. [Failure mode 2] → Mitigation: [...]

Missing risk factors:
- [Risk 1]
```

### Level 3 (Deep - 60 min)
**Add**: Mental Models + First Principles + Systems + Six Hats

Load and apply:
- `references/mental_models.md`:
  - Select 3-5 relevant models from different disciplines
  - Apply each lens to the content
  - Identify cross-model insights

- `references/first_principles.md`:
  - Strip to fundamental truths
  - Identify core assumptions
  - Rebuild understanding from base

- `references/systems_thinking.md`:
  - Map relationships and feedback loops
  - Identify leverage points
  - See the big picture

- `references/six_hats.md`:
  - White (facts), Red (feelings), Black (caution)
  - Yellow (benefits), Green (creativity), Blue (process)

```markdown
## Multi-Model Analysis

### Mental Models Applied:
1. **[Model 1 from X discipline]**
   Insight: [...]

2. **[Model 2 from Y discipline]**
   Insight: [...]

3. **[Model 3 from Z discipline]**
   Insight: [...]

Cross-model pattern: [Key insight from combining models]

### First Principles Breakdown:
Core assumptions:
1. [Assumption 1] → Valid: [Yes/No/Conditional]
2. [Assumption 2] → Valid: [Yes/No/Conditional]

Fundamental truth: [What remains after stripping assumptions]

### Systems Map:
```
[Variable A] ──reinforces──> [Variable B]
      ↑                          |
      |                          |
   balances                  reinforces
      |                          |
      └─────────<────────────────┘

Leverage point: [Where small change = big impact]
```

### Six Hats Perspective:
🤍 Facts: [Objective data]
❤️ Feelings: [Intuitive response]
🖤 Cautions: [Risks and downsides]
💛 Benefits: [Positive aspects]
💚 Ideas: [Creative alternatives]
💙 Process: [Meta-thinking]
```

### Level 4 (Research - 120 min+)
**Add**: Cross-source comparison via web_search

Use web_search to find 2-3 related sources, then:
- Load `references/comparison_matrix.md`
- Compare SCQA across sources
- Identify consensus vs. divergence
- Synthesize integrated perspective

```markdown
## Multi-Source Analysis

### Source 1: [This article]
S-C-Q-A: [Summary]
Key claim: [...]

### Source 2: [Found article]
S-C-Q-A: [Summary]
Key claim: [...]

### Source 3: [Found article]
S-C-Q-A: [Summary]
Key claim: [...]

## Synthesis

**Consensus**: [What all agree on]
**Divergence**: [Where they differ]
**Unique value**: [What each contributes]
**Integrated view**: [Your synthesis]
```

## Step 4: Synthesis & Output

**Generate based on user goal:**

### For Problem-Solving:

```markdown
## Applicable Solutions
[Extract 2-3 methods from content]

## Application Plan
Problem: [User's specific issue]
Relevant insights: [From analysis]

Action steps:
1. [Concrete action with timeline]
2. [Concrete action with timeline]
3. [Concrete action with timeline]

Success metrics: [How to measure]

## Risk Mitigation (from Inversion)
Potential failure points:
- [Point 1] → Prevent by: [...]
- [Point 2] → Prevent by: [...]
```

### For Learning:

```markdown
## Learning Notes

Core concepts (explained simply):
1. **[Concept 1]**: [Definition + Example]
2. **[Concept 2]**: [Definition + Example]

Mental models gained:
- [Model 1]: [How it works]

Connections to prior knowledge:
- [Link to something user already knows]

## Deeper Understanding (First Principles)
Fundamental question: [...]
Core principle: [...]

## Verification Questions
1. [Question to test understanding]
2. [Question to test application]
3. [Question to test evaluation]
```

### For Writing Reference:

```markdown
## Key Arguments & Evidence
[Structured extraction with page/paragraph numbers]

## Quotable Insights
"[Quote 1]" — Context: [...]
"[Quote 2]" — Context: [...]

## Critical Analysis Notes
Strengths: [For citing]
Limitations: [For balanced discussion]

## Alternative Perspectives (from Mental Models)
[What other disciplines would say about this]

## Gaps & Counterfactuals
What the article doesn't address:
- [Gap 1]
- [Gap 2]
```

### For Decision-Making:

```markdown
## Decision Framework

Options presented: [A / B / C]

Multi-model evaluation:
- Economic lens: [...]
- Risk lens (Inversion): [...]
- Systems lens: [...]

## Six Hats Decision Analysis
🤍 Facts: [Objective comparison]
🖤 Risks: [What could go wrong]
💛 Benefits: [Upside potential]
💚 Alternatives: [Other options not considered]
💙 Recommendation: [Synthesized advice]

## Scenario Analysis (from Inversion)
Best case: [...]
Worst case: [...]
Most likely: [...]
```

## Step 5: Knowledge Activation

**Always end with:**

```markdown
## 🎯 Immediate Takeaways (Top 3)

1. **[Insight 1]**
   Why it matters: [Personal relevance]
   One action: [Specific, time-bound]

2. **[Insight 2]**
   Why it matters: [Personal relevance]
   One action: [Specific, time-bound]

3. **[Insight 3]**
   Why it matters: [Personal relevance]
   One action: [Specific, time-bound]

## 💡 Quick Win
[One thing to try in next 24 hours - make it TINY and SPECIFIC]

## 🔗 Next Steps

**To deepen understanding:**
[ ] Further reading: [If relevant]
[ ] Apply framework X to topic Y
[ ] Discuss with: [Who could add perspective]

**To apply:**
[ ] Experiment: [Test in real context]
[ ] Teach: [Explain to someone else]
[ ] Combine: [Mix with another idea]

## 🧭 Thinking Models Used
[Checkboxes showing which frameworks were applied]
✅ SCQA ✅ 5W2H ✅ Critical Thinking ✅ Inversion
□ Mental Models □ First Principles □ Systems □ Six Hats
```

## Quality Standards

Every analysis must:
- ✅ Stay faithful to original content (no misrepresentation)
- ✅ Distinguish facts from opinions
- ✅ Provide concrete examples
- ✅ Apply frameworks appropriately (not force-fit)
- ✅ Connect to user's context when possible
- ✅ End with actionable steps
- ✅ Cite specific sections (paragraph numbers, quotes)

**Avoid:**
- ❌ Overwhelming with all frameworks at once (respect depth level)
- ❌ Academic jargon without explanation
- ❌ Analysis without application
- ❌ Copying text verbatim (always reword for understanding)
- ❌ Using frameworks superficially (go deep, not wide)

## Interaction Patterns

**Progressive questioning:**
- Understanding: "What do you think the author means by X?"
- Critical: "Do you see any gaps in this argument?"
- Application: "How might you use this in your work?"
- Meta: "Which thinking model helped you most? Why?"

**Adapt to signals:**
- User asks "what's the main point?" → They want conciseness, use SCQA
- User challenges your analysis → Lean into Critical Thinking + Inversion
- User asks "how do I use this?" → Focus on application + First Principles
- User wants "multiple perspectives" → Use Six Hats or Mental Models
- User mentions "risks" → Apply Inversion Thinking
- User asks "how does this connect?" → Use Systems Thinking

**Framework suggestions during conversation:**
- "Would you like me to apply [X framework] to this point?"
- "This seems like a good place for inversion thinking - want to explore failure modes?"
- "I notice several mental models at play here, want me to unpack them?"

## Reference Materials

### Core Frameworks (All Levels)
- `references/scqa_framework.md` - Structure thinking (S-C-Q-A)
- `references/5w2h_analysis.md` - Completeness check (7 questions)

### Standard Level Frameworks
- `references/critical_thinking.md` - Argument analysis
- `references/inversion_thinking.md` - Risk and failure mode analysis

### Deep Level Frameworks
- `references/mental_models.md` - Multi-discipline model library
- `references/first_principles.md` - Essence extraction method
- `references/systems_thinking.md` - Relationship mapping
- `references/six_hats.md` - Multi-perspective protocol

### Output Formats
- `references/output_templates.md` - Note format examples
- `references/comparison_matrix.md` - Cross-article analysis

## Advanced Usage

### Custom Framework Combinations

User can request specific combinations:
- "Use SCQA + Inversion" - Structure with risk analysis
- "Apply Mental Models + Systems Thinking" - Multi-lens system analysis
- "5W2H + Critical Thinking" - Completeness + quality check

### Iterative Deepening

Start with Level 1, then ask:
- "Want to go deeper on any part?"
- "Which framework would be most valuable here?"
- "Should we do an inversion analysis of this solution?"

### Domain-Specific Optimizations

**Business/Strategy**: SCQA + Mental Models (economics) + Inversion
**Technical/Research**: 5W2H + First Principles + Critical Thinking
**Personal Development**: Six Hats + Inversion + Systems
**Decision-Making**: Mental Models + Inversion + SCQA
**Creative**: Six Hats + First Principles + Mental Models

---

**Remember**: The goal is insight, not framework completion. Use frameworks as tools to reveal understanding, not as checklists to complete. Quality of thinking > quantity of frameworks applied.

© YYH211 · MIT · YYH211/Claude-meta-skill

Похожие скиллы

Все скиллы по теме «Тексты и документация» →
deep-research claude-office-skills Скилл deep-research для Claude Code ведёт исследование по теме по фазам: скоуп, сбор, анализ, синтез, отчёт. Выдаёт структурированный разбор, а не набор ссылок. CX Codex +1 5 200установок research_analyst AgenticOrgLab Субагент research_analyst для Claude Code собирает исследовательский отчёт по заданному вопросу и переданным источникам. Он сопоставляет противоречивые утверждения и отмечает выводы отдельно от фактов из источников. CX Codex ★ 776звёзд репо sociologist-analyst franklee16 Скилл sociologist-analyst для Claude Code анализирует события и группы через социологические рамки: институты, стратификацию, нормы и коллективное поведение. Разбирает социальные сюжеты. CX Codex +1 ★ 240звёзд репо now-reading-page readwiseio Скилл now-reading-page для Claude Code создаёт автономную HTML-страницу по данным Readwise Reader: показывает прогресс чтения и сортирует материалы по статусу. CX Codex +1 461установок deepread xiehuan123 Плагин deepread разбирает книги и PDF с упором на доказательства. Помогает проверить аргументы через пять режимов чтения. ★ 30 251звёзд репо dsh-deepread xiehuan123 Плагин dsh-deepread для Claude Code разбирает статьи и книги в пяти режимах — от краткого обзора до карты знаний и плана повторения. Результаты используют для проверки фактов и подготовки заметок. ★ 3 785звёзд репо

Из того же набора4

Спонсоры проекта

Компании, которые помогают держать аналитику открытой