TRUST CENTER
CyberDefender Trust Center
A public, evidence-conscious view of CyberDefender's current R&D/MVP-stage security boundaries.
Pilot-stage transparency
CyberDefender is an active engineering and R&D/MVP project preparing for pilot deployment. This page distinguishes implemented public boundaries from future or planned capabilities.
Current public security boundaries
Public websiteSeparated from the authenticated Owner plane and the machine Control Plane API.
Local originsPublic, Owner and API origins are designed to remain loopback-bound behind controlled gateway/tunnel paths rather than direct inbound exposure.
AI assistantPublic informational/advisory scope only. No Owner credentials, fleet-token access or privileged OS execution.
Commercial pilotObserve-first: monitoring, detection, correlation, risk and recommendations. Destructive autonomous actions are not part of the initial pilot.
Security design principles
- Risk is not authorization.
- AI output is not treated as privileged authority.
- High-impact response is designed around policy, Safety Core and independent verification.
- Failure and uncertainty at critical authorization boundaries should fail safely rather than silently expand authority.
- Public visitors are not routed into private Owner/Admin surfaces.
Certification status
CyberDefender does not currently claim SOC 2, ISO 27001, FedRAMP or equivalent external certification on this R&D/MVP-stage website. Any future certification claim should be published only after independent completion and evidence.
Public disclosure scope
High-level architecture and pilot controls may be published. Patent-sensitive implementation details, cryptographic material, detection content, private operational procedures, credentials and other trade-secret material are intentionally excluded from the public website.