CLAWG

Every agent you run, working from the same brain.

Stop re-explaining your context to each new session. One shared vault holds who you are, how you work and what you know — and every agent loads it from anywhere.

Open-source framework · native shared Obsidian Second Brain · master context files · shared skills, subagents, tools and APIs · per-agent identity profiles · Obsidian Sync across sessions

Every session can load shared markdown context from one vault so all agents work from the same source of truth:

• master files (user.md, environment.md, philosophy.md)
• shared capabilities (skills/, subagents/, tools/, api/, learning/)
• per-agent profile files (identity.md, AGENTS.md, soul.md)

Access your context from any Claude Code session via Obsidian Sync.

Why CLAWG

• One context runtime for all agents
• Native markdown workflow (edit in Obsidian, reuse in CLI)
• Shared tools and APIs across all agents
• Multi-agent architecture with deterministic profile loading
• Open-source and self-hostable

Quickstart

1) Install

Supported: macOS, Linux, WSL2.

Documentation

• Landing page source: landingpage/index.html
• Complete setup guide: docs/SECOND_BRAIN_SETUP.md
• Additional docs: docs/

Contributing

• Open an issue describing the change.
• Create a branch from main.
• Keep pull requests focused and testable.
• Add docs updates when changing runtime behavior.

License

MIT - see LICENSE.