Use when mastering a book's content through autonomous extraction, validation against NotebookLM, and knowledge extension
3-phase autonomous agent: extract book content → validate against NotebookLM → extend beyond the source. Inspired by autoresearch — independent teams with information barriers.
All knowledge lives in Obsidian, never in this skill.
obsidian_vault: "{vault}/knowledge/raw" # Read {vault} from ${CLAUDE_PLUGIN_ROOT}/brain-os.config.md
notebooklm_bin: ~/.local/bin/notebooklm
scripts_dir: ${CLAUDE_PLUGIN_ROOT}/skills/self-learn/scripts
Goal: Read the book → identify atomic concepts → write Obsidian notes.
scripts/lib/epub_parser.py to get chapter textscripts/lib/note_writer.py
c. Categorize into topic folders, link to related concepts from previous chapters{obsidian_vault}/{book-slug}/
├── decision-making/
├── business-operations/
├── leadership/
└── [discovered categories]/
---