Skip to content

README.md

Latest commit

 

History

History
76 lines (57 loc) · 3.84 KB

README.md

File metadata and controls

76 lines (57 loc) · 3.84 KB

deep-learning — Deep Reading for Zettelkasten

All-in-one deep reading skill: turn books, long articles, and reports into a connected knowledge network (structure notes, atomic notes, method notes, index). Uses Mortimer Adler for structure, Feynman for clarity, Luhmann for linking, plus Pragmatist and Critics. Enforces case fidelity and actionable extraction.

When to Use

Use this skill when:

  • You want to deeply digest a book, long article, research report, or paper and build a knowledge network (not just a summary).
  • You need structure notes (reading order + argument tree), atomic notes (concepts), method notes (SOPs/checklists), and index notes (entry points).
  • You say things like: “help me turn this book into a Zettelkasten,” “depth-first read this report,” “structure note + atomic notes for this article.”

Quick Reference

Phase Output Key artifact
0 Pre-game plan YYYYMMDD_01_[title]_execution_plan.md
1 Structure note templates/structure_note_template.md
2 Index note templates/index_note_template.md
2.5 Index onboarding Mount to existing index; move to index dir
3 Atomic notes + Luhmann Scan references/luhmann_scan.md
4 Method notes templates/method_note_template.md
5 Feynman review De-jargon, logic, topology
6 Network review ≥2 links per note; multi-index mount
6.5 Workflow audit (optional) If workflow-audit skill is present

Workflow at a Glance

  1. Phase 0: Produce an execution plan (≥6 TODOs + context).
  2. Phase 1: Create structure note (core thesis + logic chain). Uses structure-note if available.
  3. Phase 2: Create index note (keywords + entry points). Uses index-note if available.
  4. Phase 2.5: Mount the new index to an existing index; move index file into your index directory.
  5. Phase 3: Create atomic notes from structure; run Luhmann Scan (dependencies, links, methods) per note.
  6. Phase 4: Turn discovered methods into method notes (SOP/template/checklist).
  7. Phase 5: Feynman review (clarity, metaphor, logic, surprise links).
  8. Phase 6: Ensure ≥2 bidirectional links per note; complete index onboarding; multi-index mount check.
  9. Phase 6.5: Optional workflow-audit pass if that skill is installed.

Requirements

  • Paths: The skill assumes an index directory (e.g. Index/) and a daily/task directory (e.g. Daily/YYYY/MM/DD). Adapt path names in the skill or your vault to match.
  • Optional companion skills: For full automation, structure-note, index-note, file-organize, and workflow-audit can be used when available; the skill still works with manual steps if they are missing.

Install (skills.sh)

If this skill is published to GitHub:

npx skills add <owner/repo>

Example (when repo is your-username/deep-learning):

npx skills add your-username/deep-learning

Repo Layout

deep-learning/
├── README.md           # This file
├── SKILL.md            # Main skill instructions (required)
├── references/
│   ├── expert_personas.md
│   └── luhmann_scan.md
└── templates/
    ├── atomic_note_template.md
    ├── index_note_template.md
    ├── method_note_template.md
    └── structure_note_template.md

Publishing to GitHub and skills.sh

See PUBLISH.md for step-by-step: create GitHub repo, push this folder, add description/topics, then install via npx skills add <owner>/deep-learning and optionally submit on skills.sh.