Skip to content

Augment Claude with this skill to help create atomic flashcards from a bank of sources (project ideally) that help you learn based precisely on the science.

Notifications You must be signed in to change notification settings

jalliet/flashcards

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 

Repository files navigation

Generating STEM Flashcards

A Claude skill for generating atomic, cognitively-principled flashcards from technical source material.

What It Does

Upload lecture notes, textbooks, or technical documentation. Get interactive flashcards organised into three cognitive layers:

Layer Tests Example
L1: Recall Facts, definitions, formulas "Define the Jacobian matrix"
L2: Understanding Why/how, intuitions "Why is the Jacobian useful for coordinate transforms?"
L3: Boundaries Limitations, edge cases "When does the Jacobian become singular?"

The skill enforces atomicity (one concept per card), refuses to card inappropriate content (proofs, worked examples), and renders mathematical notation via KaTeX.


Installation

  1. Clone this repo or download as ZIP: git clone https://github.com/jalliet/flashcards.git
  2. Create a skill package: zip -r flashcards.zip SKILL.md references/ assets/
  3. Upload to Claude.ai: Settings > Capabilities > Skills > Upload Skill

Usage

Basic

Upload a PDF, paste lecture notes, or describe a topic:

"Generate flashcards spanning topicas from all the project files."

"Create flashcards covering gradient descent, including failure modes"

"Make revision cards for chapters 3-5 of the uploaded PDF"

With Layer Filtering

"Generate only L3 (boundary) cards for this material"

"Focus on L1 recall cards for definitions and formulas"

Iterative Refinement

"Card 12's formula is missing the inverse; fix it"

"Split card 7 into two separate cards"

"Add more L2 cards for the eigenvalue section"


Recommended Setup

Projects

Create a Claude Project for your course/subject. Add:

  1. This skill (.skill file)
  2. Your lecture notes and readings
  3. Project instructions like: "Always use the flashcard skill when I ask for revision materials"

Complementary Skills

For best results with complex documents, also install from anthropics/skills:

Skill Why
pdf Better extraction from scanned/complex PDFs
docx Preserves formatting from Word documents
frontend-design Improves React artifact rendering and styling for academic content

The flashcard skill works without these, but they improve source parsing.


Output

The skill produces an interactive React artifact with:

  • Layer filtering (L1/L2/L3 tabs)
  • Topic filtering
  • Star/favourite cards
  • Shuffle mode
  • Responsive design

Cards can be copied to Anki or other spaced repetition systems.


File Structure

generating-stem-flashcards/
├── README.md                 # This file
├── THEORY.md                 # Scientific foundations (Bloom, CLT, etc.)
├── SKILL.md                  # Main skill instructions
├── references/
│   ├── COGNITIVE_LAYERS.md   # L1/L2/L3 definitions
│   ├── ATOMICITY.md          # Quality rules, refusal policy
│   └── LATEX_SYNTAX.md       # KaTeX reference
└── assets/
    └── flashcard_template.jsx

Note: Only the skill files (SKILL.md, references/, assets/) are packaged in the .skill file. README.md and THEORY.md are GitHub documentation only.


Why This Skill?

Most AI flashcard generators produce low-quality cards: compound questions, no cognitive framework, no quality control.

This skill encodes learning science directly into generation:

  • Bloom's Taxonomy → Three-layer structure
  • Cognitive Load Theory → Atomicity rules
  • Minimum Information Principle → Refusal policy

See THEORY.md for the full scientific foundation with citations.


License

MIT

About

Augment Claude with this skill to help create atomic flashcards from a bank of sources (project ideally) that help you learn based precisely on the science.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published