What it does
open-brain-extract takes any source — a PDF, an article, a Twitter thread, a transcript — and converts it into atomic Literature Notes inside your Obsidian vault.
Each resulting note has:
- Structured YAML frontmatter with source, layer, domain, and search_terms
- Automatic classification into one of 7 knowledge layers
- “Your take” placeholders for you to add your own perspective
- Candidate links to existing notes in the vault
The 7 layers
- Facts — verifiable data points, statistics
- Concepts — definitions, mental models
- Procedures — step-by-step how-tos
- Context — when and where it applies
- Relations — connections between ideas
- Metacognition — how to think about the topic
- Transfer — applications to other domains
How to use it
Inside Claude (with the skill installed in Cowork or Claude Code):
extract notes from this PDF: [attach file]
The skill handles the rest.
Example output
---
title: "Memory consolidation in LLMs requires spaced repetition"
source: "MemoryConsolidation2025.pdf"
layer: metacognition
domain: AI/Memory
search_terms: [memory consolidation, spaced repetition, LLM, retrieval practice]
---
Memory systems in LLMs show patterns analogous to human consolidation:
spaced repetition of contexts improves factual retention
across long context windows.
**Your take**: ...