Governance
A platform is not complete until roles, audit, boundaries, and versioning are explicit.
This layer keeps the engine adoptable without turning vague: workspaces, role separation, white-label boundaries, profile-specific rules, audit visibility, and conformance versioning.
Governance block
What it controls
Why it exists
Workspace roles
Owner, analyst, reviewer, partner admin
Stops one workspace from becoming a shared uncontrolled inbox.
Branding mode
Standard, co-branded, partner-primary
Lets white-label adoption exist without hiding the engine boundary.
Profile lock
Which modules and reason-code paths are active
Prevents generic drift across verticals.
Audit trail
Creation, scan, simulation, export, update events
Makes partner and platform operations reviewable.
Version policy
Schema, registries, vectors, conformance
Keeps integrations stable as the engine evolves.
Minimum governance contract
- One workspace = one profile posture
- Every scan record gets a case ID and timestamp
- Every reviewer simulation is persisted or explicitly discarded
- White-label presets may change brand shell, not engine facts
- Conformance vectors must keep passing before profile updates ship