Shared Brain

AI agents that learn from each other’s mistakes — and prove it.

Shared Brain is a CLI tool that checks commands against structured lessons before execution, logs compliance, and builds audit trails.

Quick Start

pip install shared-brain
brain hook install

That’s it. Every Bash command in Claude Code now gets checked against your lessons.

Why?

Rules don’t work. We wrote “Always GET before PUT” in our instructions. The AI agent ignored it. Twice. Five articles deleted.

Shared Brain turns rules into enforced guardrails:

$ brain guard "curl -X PUT https://api.example.com/articles/123"

⚠️  CRITICAL: api-put-safety (violated 2x)
   PUT replaces the ENTIRE resource.

   Checklist:
   [ ] GET the current resource state
   [ ] PUT body contains ALL required fields

Proceed? [y/N]

Features

Guard: Pattern-match commands against YAML lessons
Audit: JSONL trail of every check — who, what, when, followed?
Stats: Compliance rates per lesson, per agent
Hook: One-command Claude Code integration
Zero deps: Pure Python stdlib. No PyYAML needed.

Links