How Our AI Brain System Works

A single file that tells your AI who it is, how to behave, what rules to follow, and how your project works. Think of it as your AI's instruction manual — its identity, personality, and operating procedures in one place.

Skills are instruction files that teach your AI specific capabilities. Each skill is like an app — it gives your AI knowledge about a particular domain or task.

Agents are specialized AI roles with persistent memory. Each agent focuses on one area of expertise and remembers context across sessions. They work together as a team — invoking each other through shared memory files and passing context via the memory system. The orchestration pattern: you trigger a command → the command routes to the right specialist → the specialist writes results to shared state for others to build on.

Slash commands are pre-built workflows you trigger with a simple command. Type /start to begin your day, /audit to review quality, /plan to architect a solution.

Hooks are automated behaviors that run before or after specific actions. They enforce quality gates and safety checks without you having to think about them.

The system is designed with deliberate human oversight built into the architecture — not bolted on as an afterthought. Before knowledge gets promoted to the permanent knowledge base, the Auditor agent reviews it. Before destructive operations run, hooks enforce safety checks. This is an architectural safety pattern: the AI cannot silently modify its own rules or take irreversible actions without a checkpoint.

A 7-tier memory architecture that ensures your AI retains knowledge across sessions. From short-term working memory to long-term learned rules — nothing is lost.