Punk Dev Robot Dotfiles Generate Anki Deck

Generate Anki Deck

Usage: generate-anki-deck [convert|generate|build|full <source-path>] [--chapters N,N-N]

Step 1: Determine Workflow

Infer the workflow from provided arguments. If none provided, use AskUserQuestion:

"What would you like to do?"

Options:

1. Convert — source (PDF/document) to markdown chapters
2. Generate — Anki cards from existing chapters
3. Build — .apkg deck from existing card JSON
4. Full pipeline — source → chapters → cards → deck

If --chapters is provided (e.g., --chapters 3, --chapters 3,5-7), pass the selection through to the generate step — skip the "which chapters?" prompt.

Step 2: Load Workflow

Read the workflow file for the chosen step:

Then execute the steps in that workflow file.

Card Quality Principles

These apply to ALL card generation — non-negotiable.

1. Atomic concepts: Each card tests ONE idea. Never combine multiple concepts.
2. Concise questions: Clear and direct, 1-2 sentences max.
3. Detailed answers: Explain WHY, not just WHAT. Include context and implications.
4. ASCII diagrams: For architecture or flow content, include simple ASCII diagrams in answers (box-and-arrow, under 10 lines).
5. No trivial facts: Skip page numbers, author names, copyright info, "see chapter X" references.
6. No multi-concept cards: If testing two ideas, make two cards.

Output Structure

All output goes to ./anki-output/ in the current working directory, scoped by source:

./anki-output/{source-slug}/
├── chapters/              # 01-chapter-title.md, ...
├── cards/                 # chapter-01.json, ...
└── {source-slug}.apkg     # Final deck

The source slug is derived from the filename or title (kebab-case, lowercase).

Constraints

• Do not enter plan mode. Execute workflow steps directly.
• Scripts are executables. Run them directly (e.g., scripts/extract_pdf.py). Do not read their source code.
• Load files on demand. Only read the workflow file for the current step. Do not pre-load all workflows or references.
• Do not write custom Python. All PDF extraction and deck building is handled by the provided scripts.

Converted and distributed by TomeVault — claim your Tome and manage your conversions.