STARTMAKINGSENSE

Interoperability Standards

Standards and protocols that define how identity-aware AI systems interoperate across IAAI interfaces.

MCP

Commons DraftOpen Standard

Steward: Linux Foundation

Inter-pillar interfaces where active: A-B

Enables centralized policy enforcement and SSO-anchored identity for AI agent tool access. Client-server protocol for AI applications connecting to external tools, data sources, and services.

MCP Enterprise-Managed Authorization

Commons DraftOpen Standard

Steward: Linux Foundation

Inter-pillar interfaces where active: A-B

Extension of MCP that routes MCP server authorization through enterprise IdP using OAuth 2.1 scoped tokens, enabling centralized governance for AI agent tool access.

Microsoft Graph API

Commons DraftVendor Standard

Steward: Microsoft

Inter-pillar interfaces where active: B-C

API layer enabling Copilot and abstraction-layer services to retrieve Entra group membership and sensitivity label scope for identity-aware output gating.

NIST AI RMF

Commons DraftOpen Standard

Steward: NIST

Inter-pillar interfaces where active: A-E, D-E

Aligns Pillar E governance function with NIST RMF Govern function; Map/Measure functions align with use case triage and security findings.

NIST SP 800-207

Commons DraftOpen Standard

Steward: NIST

Inter-pillar interfaces where active: A-B, A-C, B-D

IAAI five-pillar architecture implements ZTA for AI systems with Pillar A as policy engine, B/C as enforcement points, D as continuous monitoring.

OAuth 2.0 RFC 8693

Commons DraftOpen Standard

Steward: IETF

Inter-pillar interfaces where active: A-B, B-C

Formal standard for propagating user identity through AI agent call chains with complete auditable delegation. Token exchange for delegation chains (On-Behalf-Of pattern); resulting token carries subject (originating human) and actor (delegating agent) claims.

OPA REST API

Commons DraftOpen Standard

Steward: Cloud Native Computing Foundation (CNCF) / Open Policy Agent

Inter-pillar interfaces where active: A-B

Exposes Open Policy Agent policy evaluation as a REST endpoint, enabling Pillar A policy enforcement for AI agent authorization decisions.

SCIM 2.0

Commons DraftOpen Standard

Steward: IETF (RFC 7643/7644)

Inter-pillar interfaces where active: A-B, A-E

Wire protocol enabling IGA platforms to push identity lifecycle events to AI-touching systems automatically. Draft IETF extension adds Agent and AgenticApplication resource types.

SPIFFE

Commons DraftOpen Standard

Steward: Cloud Native Computing Foundation (CNCF)

Inter-pillar interfaces where active: A-B, B-D

Eliminates long-lived secrets for AI agents; enables mTLS between services with automatically rotated certificates using X.509 SVIDs (SPIFFE Verifiable Identity Documents).