AUVIRA
Engine Healthy

Admin

Settings

Control-plane defaults applied to every AI system unless overridden at the system level.

Enforcement Defaults

Default enforcement mode

Decisions are applied inline; disallowed responses and actions are blocked.

Fail closed

If the control plane cannot reach a decision, refuse the request rather than allowing it through.

Redact PII in prompts

Mask detected personal and credential data before it reaches the foundation model.

Gateway & Data

Mirror mode for new models

Newly connected models start in mirror (observe-only) interception until promoted to inline.

Decision provenance retention

Immutable evidence records are kept for 365 days before cryptographic expiry.

API Credentials

Live enforcement key

vrd_live_9f2c8a1e4b7d6033c15e2a90

Requests authenticate with this key at the edge before any policy is evaluated.

Integrations

Splunk SIEM

Stream decision events for correlation

Slack

Route human-review escalations to a channel

PagerDuty

Page on critical hard-stop decisions

ServiceNow GRC

Sync policy versions and evidence

Team

AO

a.okonkwo@corp.io

Admin
AR

a.rahman@corp.io

Policy Author
MC

m.chen@corp.io

Reviewer
SN

s.novak@corp.io

Reviewer