Security

Set boundaries for agent actions.

Tenant isolation, governed tools, network restrictions, signed reports, and SLO-backed release evidence are built into the platform surface.

Forced row-level security for tenant-scoped data.

Policy gates for risky tool execution.

Private endpoint and platform secret protections.

Audit trails across auth, tools, workflows, and releases.

4

Core workspaces

4

Agent modes

AA

Accessibility target

0-3

Risk levels