Claude Code + Obsidian The AI Second Brain Playbook

You open Claude Code. It forgets what you decided yesterday. It doesn't know your project structure. It doesn't remember the 12 trade-offs you already reasoned through. Again.



Meanwhile, developers on reddit are running 285-line operating manuals inside Obsidian vaults with 5 lifecycle hooks that load context before they even type. The difference isn't talent. It's a vault Claude actually reads.



This book gives you the complete Karpathy-method system, productized. You'll ship a working vault on Saturday. By Sunday, your CLAUDE.md sticks. By next weekend, you'll have 5 lifecycle hooks, 5 skills, and 3 subagents loaded. Claude stops asking "what were we doing?" By day 30, your vault has a role-specific weekly review and a semantic search layer that finds notes by concept, not keyword.



This isn't another "intro to Obsidian" book. The 7 direct competitors all have zero customer reviews. One is 28 pages of filler. One is 563 pages of theory at $34.99 print. Three don't name Claude Code in the title. None ship the CLAUDE.md template that Claude actually follows. None walk through the 5-hook pattern. None give you the subagent definitions.



Here's what you'll build:


1. The 5-folder vault structure Claude can actually use.


2. A CLAUDE.md that Claude follows on every session, not just the first.


3. Five lifecycle hooks that fire before, during, and after every prompt.


4. Five named skills: /capture, /daily, /plan, /process-inbox, /weekly-review.


5. Three subagents: brag-spotter, slack-archaeologist, cross-linker.


6. A role-specific weekly review for PMs, founders, freelancers, or knowledge workers.


7. A semantic search layer that finds notes by concept, even past 2,000 notes.



This is 170+ pages of dense, copy-paste-ready playbooks and examples. Scroll up and grab your copy.