- This document proposes best practices for authentication and authorization of AI agents by leveraging existing standards (e.g., WIMSE, OAuth 2.0, SPIFFE) rather than defining new protocols.
- AI agents are defined as workloads that interact with LLMs, tools, services, and resources; they require unique identities, short-lived cryptographic credentials, and runtime credential provisioning with posture assessment.
- The Agent Identity Management System (AIMS) is a conceptual model comprising identifiers, credentials, provisioning, authentication, authorization, observability/remediation, policy, and compliance.
- Transport-layer authentication (mTLS) and application-layer authentication (WIMSE Proof Tokens, HTTP Message Signatures) are both supported, with application-layer preferred when intermediaries break transport continuity.
- Authorization leverages OAuth 2.0 flows (e.g., token exchange, client credentials, delegation chaining) with standards like UMA, AuthZEN, and CIDPA for policy enforcement.
- Monitoring and observability are security controls; deployments must produce tamper-evident audit logs and can use OpenID Shared Signals Framework for real-time remediation.
- Security best practices include OAuth 2.0 BCP and FAPI 2.0 profiles; static API keys are considered an antipattern.