Skillget

learn-anything

agent
Security Audit
Fail
Health Warn
  • License — License: MIT
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Low visibility — Only 5 GitHub stars
Code Fail
  • rm -rf — Recursive force deletion command in tests/test_epub.sh
Permissions Pass
  • Permissions — No dangerous permissions requested

No AI report is available for this listing yet.

SUMMARY

an agent skill generates good material to learn anything

README.md

Learn Anything

Structured learning framework for any subject — study via CLI with spaced repetition (SM-2). Three-theory pedagogy: Marva Collins, Feynman Technique, Desirable Difficulties.

Features

  • Interactive curriculum creation: LLM-driven syllabus + lesson + MCQ generation
  • CLI study system: read lessons, Feynman explain-back, MCQ drills, spaced repetition
  • SM-2 spaced repetition: optimal recall intervals, mixed-module interleaving
  • Cost-effective: ~$0.10 per course creation, $0 per study session
  • Portable: Markdown files importable into Anki, Obsidian, Notion

Quick Start

  1. Trigger creation: Say I want to learn [topic] or learn.sh init <subject>
  2. Study: learn.sh start <subject> → learn.sh quiz <subject> <module>
  3. Review daily: learn.sh review <subject>

CLI Commands

learn.sh init <subject> [lang]   Create new subject (en|zh|yue)
learn.sh start <subject>         Show overview + modules
learn.sh quiz <subject> <mod>    MCQ drill
learn.sh explain <subject> <mod> Feynman technique prompt
learn.sh feynman <subject> <mod> Alias for explain
learn.sh review <subject>        SM-2 spaced repetition
learn.sh stats <subject>         Progress + retention
learn.sh export <subject>        Anki CSV export
learn.sh epub <subject> [file]   Export course to EPUB book

Subject Structure

subjects/<topic>/
├── syllabus.yaml       # Course spec
├── modules/
│   ├── NN-name/
│   │   ├── lesson.md   # Core content, examples, reframe prompts
│   │   └── quiz.yaml   # 8-10 MCQs
│   └── ...
└── srs/
    ├── deck.json       # SM-2 card state
    └── stats.json      # Study history

Pedagogy

Theory Role
Marva Collins Rigor, repetition, high expectations. Read thoroughly, answer precisely.
Feynman Technique Explain simply → find gaps → refine. Teach concept to imaginary child.
Desirable Difficulties Spaced repetition, interleaved modules, varied MCQ difficulty.

Session Types

Session Duration Focus
LEARN 45-60 min Read lesson, reframe, MCQ drill
EXPLAIN 15-20 min Feynman explain-back, AI gap detection
REVIEW 10-15 min SM-2 spaced repetition (daily)
MIXED 30-45 min REVIEW + LEARN + EXPLAIN combined

EPUB Generation

After creating course content, generate portable EPUB book:

learn.sh epub <subject>              # Full build: assemble + generate
learn.sh epub-regen <subject>        # Regenerate from cached book.md (faster)
learn.sh epub-verify <subject>       # Validate EPUB structure

Agent workflow

  1. Create all modules via content creation protocol (Section 3)
  2. Run learn.sh epub <subject> to assemble lessons + quizzes into single EPUB
    • Script collects all lesson.md + quiz.yaml → writes book.md → generates EPUB
  3. Run learn.sh epub-verify <subject> to validate output
  4. To update after module edits: learn.sh epub-regen <subject> (skips assembly)

Manual alternative: epub.py build <subject-dir> <output> or epub.py from-md <book.md> <output>.

Cost Model

Powered by DeepSeek V4 Flash.

Phase Cost
Scope + syllabus ~$0.01
Per module (~15K tokens) ~$0.004
Full course (20 modules) ~$0.08
Per study session $0

License

MIT

Reviews (0)

No results found