Browse / General-purpose / big-brain

big-brain

marcogalluccio

archived

An open-source operational workspace template powered by Claude Code

Open marcogalluccio/big-brain → Checked against the four rules.

Built for one agent

ships CLAUDE.md and nothing for other agents.

In its own words

README.mdmarcogalluccio/big-brain

⚠️ This template has evolved into Big Brain OS — a personal operating system for your work, with two-level memory, self-improving skills, and Superpowers-first onboarding. This repository is archived and kept for reference; all new development happens there.

An open-source operational workspace template powered by Claude Code.

Big Brain gives you a structured system where AI agents understand your full context — who you are, what your organization does, and how your work is organized — before they do anything.

The Three Core Ideas

1. CLAUDE.md Hierarchy

Every folder has a CLAUDE.md file that tells Claude Code what that area of the workspace is about. Files load top-down: the root CLAUDE.md sets global rules, subfolder CLAUDE.md files add domain-specific context.

The agent reads the root CLAUDE.md on every conversation. It reads subfolder CLAUDE.md files only when working in that area. This keeps context focused and avoids noise.

2. File-Based Memory System

Memory lives in ~/.claude/projects/[your-workspace]/memory/. It persists across conversations and is the source of truth for what is active, what is planned, and what has been decided.

MEMORY.md is loaded automatically by Claude Code. Individual files hold full context and are read on demand. The daily briefing skill reads MEMORY.md to build your priorities each morning. The session debrief skill updates it at end of day.

Quoting 57% of the README. Read the rest on GitHub ↗

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