Browse / Second brain / second-brain

second-brain

cdcoonce

A graph-first Obsidian + Claude Code 'second brain' starter — skills, hooks, and templates for a persistent, self-maintaining personal knowledge vault.

Open cdcoonce/second-brain → Manually confirmed as a workspace.

Built for Claude Code

ships .claude CLAUDE.md and nothing for other agents.

In its own words

README.mdcdcoonce/second-brain

A starter template for a persistent, graph-first "second brain" you drive with Obsidian and Claude Code.

The problem it solves: an AI assistant starts every session blank, and your raw thoughts, meeting notes, incidents, and ideas need a home that's searchable, linked, and still useful months later. This vault is that home — and it doubles as persistent memory for Claude across sessions and machines.

It is graph-first, not folder-first: folders let you browse, but [[wikilinks]] are how you think. Organization comes from the connections between notes, not from which directory they sit in. A note without links is a bug.

The four pillars

Everything here serves one of four Cs:

  • Context — what the brain knows. brain/ (North Star, Memories, Key Decisions, Patterns,

Gotchas), workspace notes, and enforced frontmatter. Context is a contract, not a convention.

  • Connections — how knowledge links. A real bidirectional wikilink graph, with backlinks as

evidence trails. This is the crown jewel — not folders-with-search.

  • Capabilities — what the brain can do. Slash-command skills, reviewer subagents, and a

hook/automation layer that validates writes and auto-syncs.

  • Cadence — the rhythm that keeps it alive. Session-start context loading, /standup, a rolling

handoff, and weekly task rollover. A second brain that isn't maintained decays into an archive.

Quoting 34% of the README. Read the rest on GitHub ↗

Read from the repository's git tree on 2026-08-17.