Skip to main content
The security posture score gives you a single number (0-100) that reflects how well your organization governs its AI agent ecosystem. It is separate from the EU AI Act compliance readiness score - this score focuses on operational security, not regulatory compliance. Requires the admin scope.

Score Dimensions

Get Security Posture

There is no separate agent_count field and no dimensions object keyed by name - components are a flat list, each with name, max_score, score, and detail. Quarantine Status and Kill Switch Status are simple current-state counts (agents quarantined right now / agents blocked right now) - there is no “reviewed within 7 days” or “tested in the last 30 days” tracking behind these scores.

Score Ranges

Security Posture vs Compliance Readiness

Both scores are complementary. A high security posture score means your agents are well-governed right now. A high compliance readiness score means you can demonstrate regulatory compliance.

Improving Your Score

  1. Agent Ownership - Assign owners to all agents using ABAC.
  2. Agent Expiry - Renew or decommission agents whose expires_at has passed via Agent Lifecycle.
  3. Access Reviews - Configure review schedules and complete overdue reviews.
  4. Quarantine Status - Resolve quarantined agents - release or permanently block them.
  5. Kill Switch Status - Unblock agents once the underlying issue is resolved; this score reflects how many agents are blocked right now.
  6. Risk Classification - Classify all agents using Risk Classification.