AI Agent Security and Governance Preview
Agent Security is a Preview capability. It is available for evaluation and testing but is not yet production-ready. Features and APIs may change.
Why agent governance matters
AI agents increasingly operate autonomously — calling tools, delegating to sub-agents, and making decisions on behalf of users. Organizations need visibility into agent actions and evidence that those actions comply with policy.
Preview capabilities
- Actor type tracking — distinguish between human, agent, and service request origins in audit records
- Agent identity metadata — capture agent name, version, and delegation chain
- Tool access policy findings — evaluate agent tool usage against organizational policy
- Delegation audit evidence — record agent-to-agent delegation chains for compliance
Current limitations
- Agent Security is Preview — not production-ready
- Agent identification relies on client-provided metadata
- Tool access policies produce findings, not enforced blocks
- Features and APIs may change in future releases
SAIG provides runtime controls, policy enforcement, audit evidence, and compliance-supporting workflows. It does not constitute legal advice, certification, or a guarantee of regulatory compliance.
Frequently Asked Questions
Is Agent Security production-ready?
No. Agent Security is available as a Preview capability for evaluation and testing. It is not yet production-ready and features may change.
How does SAIG identify AI agents?
SAIG relies on client-provided metadata to identify actor type (human, agent, service), agent name, version, and delegation chain.
Can SAIG block agent tool access?
Currently, SAIG evaluates agent tool usage against policy and records findings. Enforcement as blocking is not yet available in the Preview.