Auto-Remediation Center

AI proposes · SOAR verifies · Policy decides · Human approves · Audit documents

Auto-Containment

Requesting a containment drafts it through the approval gate — it never executes live. Under firewall enforcement, capabilities must be enabled (none enabled yet).

Alert Triage — recent verdicts

Value Matrix — autonomy governor

recommends · never executes

Cost/benefit scoring for an action at a given threat picture. This is the same math that tightens the autonomy ladder at execution — made inspectable here.

Action

AI confidence0.80

How sure we are the threat is real (triage verdict / correlation).

Threat severity0.50

How bad if real — amplifies the stakes of acting vs. waiting.

Time sensitivity0.50

How fast it worsens — raises the cost of holding for approval.

Adjust the inputs and press Score to see the recommendation.

Default mode: narrative_plan_only · Live execution disabled · P0 actions blocked · All actions audited · P1 requires explicit approval

Loading…