uber.comwriteupSolving the Identity Crisis for AI Agents (Uber)
Uber's engineering team describes how it extended its Zero Trust stack to give AI agents verifiable identities. Each agent workload gets a SPIRE SVID, and a Security Token Service checks an Agent Registry and mints single-hop, short-lived JWTs. Those tokens carry the full actor chain (for example user → Oncall Agent → Investigation Agent), loosely modelled on OAuth 2.0 Token Exchange (RFC 8693). The tokens are enforced at an MCP Gateway for tool calls, and an AI Gateway fronts model calls with guardrails. The post also covers an A2A client in the SDK that makes the secure path the default, observability of actor chains, and a 2026 roadmap toward dynamic access control and unified policy enforcement. It is a concrete production reference for agent attribution and delegated authorization.
Organisations named here
- Uber
- OpenAI
- Anthropic
- The Linux Foundation
- IETF
- draft-klrc-aiagent-auth-01
- IETF WIMSE
- Model Context Protocol
- RFC 8693
- SPIFFE
Technologies named here
- Michelangelo
- Uber Agent SDK
- Agent Registry
- AI Agent Mesh
- Security Token Service
- MCP Gateway
- AI Gateway
- AI Guard
- SPIRE
- Standardized A2A Client
- A2A protocol
- Model Context Protocol
- agentgateway
- Kubernetes
- JWT
- OAuth
- Python
- SPIFFE
Techniques named here
- Per-hop token exchange with actor chain
- Agent identity via workload SVID and registry verification
- Single-hop, short-lived audience-bound tokens
- Tool-level policy enforcement at MCP Gateway
Related in agents
