The first brain that can say "No" (a refusal-capable brain any robot can mount)

01

Boundary Gate — five dimensions

Every action passes the boundary gate first. Failure on any one dimension triggers the refusal path.

• 01Permission — is the caller authorised for this action
• 02Environment — are the current sensor observations trustworthy, are they outliers
• 03Human safety — does the action introduce a risk of human harm
• 04Equipment safety — does the action risk damage to the equipment or workpiece
• 05Task constraints — does the action violate the task specification or contextual rules

02

Refusal four-tuple

Refusal must come with an auditable account, not silence.

• 01Reason — which dimension triggered
• 02Triggered clause — citation of the specific rule and version
• 03Alternative action — a fallback or sub-action the system can offer
• 04Escalation path — whose authorisation is required and how to appeal

03

Moat

Provable safe-refusal mechanism + auditable logs. A platform-grade standard part — any robotics vendor can integrate.