Guardian is designed around user consent, human oversight, access controls and clear separation between current capabilities and future R&D.
Preventive monitoring is optional and begins only when the user requests it and an authorised Safety Team member accepts the session.
Monitoring access is intended only for the active safety session and authorised operational roles.
Authenticated access, strict permissions, validation and role-based controls protect platform data and operations.
Safety Team members assess context and follow documented escalation workflows rather than treating every missed check-in as a confirmed emergency.
Guardian is a working MVP in active testing. We describe only what exists today; future capabilities are announced when they are ready, never presented as live.