# Extract Wisdom

> Content-adaptive wisdom extraction from text/podcast/video/article/paper. Detects wisdom domains in the content and builds custom sections rather than static headers. 5 depth levels from Instant (1 section) to Comprehensive (10-15 sections + themes). [WHAT] Port of the Miessler ExtractWisdom pattern. Spicy/contrarian takes are required. Voice follows the user's conversational style — bullets should read like "tell me what you just saw", not a press release. [WHEN] Use when: extract wisdom, analyse podcast/video/article, key takeaways, what's interesting here, distill, summarise interview, "extract wisdom from". NOT for: knowledge archiving (use archive-to-vault), atomic notes (use atomize-zettel). [LANGUAGE] Source-matching by default; English if explicitly requested.

- Skill: `carlheath/extract-wisdom` (Agent Skill)
- Install (CLI): `npx skillmds@latest add carlheath/extract-wisdom`
- Raw SKILL.md: https://api.skillmd.com/api/skills/carlheath/extract-wisdom/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Research & Search
- Author: carlheath (https://skillmd.com/u/carlheath)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/carlheath/extract-wisdom

---


# Extract Wisdom

**Role:** content-adaptive wisdom extraction.

## 5 depth levels

| Level | Sections | Use |
|-------|----------|-----|
| **Instant** | 1 (One-Sentence Takeaway) | Quickly capture main point |
| **Fast** | 3 (Top 3 + Takeaway + 2-min) | For quick-scan of an article |
| **Basic** | 3 + takeaway section | Standard for blog posts |
| **Full** (default) | 5-12 dynamic sections | Default depth for podcast/video/paper |
| **Comprehensive** | 10-15+ + Themes & Connections | Book or longer work |

## Dynamic sections

Build sections based on what the content actually is. Examples:

- **Security talk** → "Threat Model Insights" + "Defense Strategies"
- **Business podcast** → "Contrarian Business Takes" + "Money Philosophy"
- **Democracy discussion** → "Power analyses" + "Practical resistance tools"
- **PhD paper** → "Methodological insight" + "Empirical findings" + "Implications"

## Mandatory output sections (always)

- **One-Sentence Takeaway** — the single sentence anyone should remember
- **If you only have 2 minutes** — top 3 points
- **Sources & rabbit holes** — referenced material + further reading

## Spicy/contrarian takes

**Mandatory:** controversial or against-the-grain statements must NOT be softened. The user values sharp formulations. If the content had a spicy take, keep it.

## Workflow

### 1. Identify content type

```bash
# YouTube via MCP
mcp__youtube-transcript__get_transcript URL

# Web article via Exa or WebFetch
WebFetch URL

# Local file
Read path/to/file.md
```

### 2. Scan content for wisdom domains

Classify: is this security / tech / business / policy / academic / personal development / history / philosophy? Build sections accordingly.

### 3. Extract per domain

Per dynamic section, find:
- Aphorisms (citable lines)
- Core claims (theses with arguments)
- Quotes (direct quotes with author/source)
- Counter-points (what is being rejected/argued against)

### 4. Output

Markdown file with dynamic headers + mandatory sections. The user chooses:
- Save as-is to `Cards/` via atomize-zettel (for user-authored text)
- Save aphorisms to a personal WISDOM.md
- Save aphorisms to the aphorism skill database (if quoting external persons)
- Send to archive-to-vault if the whole piece is valuable

## User customization

**Tone in bullets:**
- Write as you would talk to a friend, not as a press release
- "He says that X" not "The author argues that X"
- Keep specific numbers/names
- No AI filler patterns ("notably", "key takeaway:")

**Source language:** matches the source. English podcast → English bullets, etc.

**Output language:** matches source by default. English if explicitly requested.

## Example uses

- "Extract wisdom from this podcast: [URL]"
- "What's interesting in the essay I linked?"
- "Analyse this report at Comprehensive level"
- "Quick-scan of this blog post"

## Version history

- v3.0 (2026-05-02): initial public release.