Intent Kept Next to the Change

Safe Surgery on System Files · Shipped · workstation

The “Why Brain” captures the human rationale alongside config changes and restores it into future conversation context.

You moved SSH to port 2222 on July 14th because the auth log was filling with scans. Halbert keeps the reason beside the change, with the evidence, so six months from now neither of you has to guess. The Why Brain UI is a clickable brain icon on discovery pages (GPU, Services, Storage, Containers, Network).

Clicking the brain icon records or shows a user-authored explanation for why an item is configured a certain way. Rationales are served by get_self_knowledge() and injected into future chat context, so the agent remembers your reasoning when you ask about that configuration again.

Limits and invariants

Rationales are user-curated, never auto-generated — the agent may suggest a reason, but the user confirms it.

Where this lives

• halbert_core/halbert_core/dashboard/frontend/src/components/ui/why-brain.tsx
• halbert_core/halbert_core/dashboard/routes/settings.py