TechPrufer
Sign inGet started

Blogs›Uber›AI Agent Identity and Provenance

AI Agent Identity and Provenance

AI Agent Identity and Provenance

3
posts
2025–2026

Uber is building a robust identity and access management system for AI agents to ensure accountability, auditability, and security as agentic autonomy increases. This involves establishing verifiable cryptographic identities for agents, propagating execution context across agent hops, and integrating with existing security infrastructure. The goal is to provide clear attribution for agent actions, enabling fine-grained access control and comprehensive auditing, ultimately fostering trust in AI-driven workflows.

2026

Solving the Identity Crisis for AI Agents

5/21/2026

This post details the architectural updates to Uber's identity and access technology stack to accommodate AI agents. Key components include the Agent Registry, AI Agent Mesh, Security Token Service (STS), MCP Gateway, and AI Gateway. The post outlines the process of minting JWT tokens for agents, emphasizing single-hop, short-lived tokens and full contextual attribution through the attested actor chain. It also describes the Michelangelo AI platform's options for agent development (code and no-code) and the integration of agent identity into the SDK.

How Uber Built an Agentic System to Automate Design Specs in Minutes

3/11/2026

This post introduces 'uSpec', an agentic system that uses AI agents and the Figma Console MCP to automate the generation of design component specifications. It addresses the enterprise-scale problem of manual, time-consuming, and error-prone spec writing by enabling AI agents to crawl Figma files, extract component data (anatomy, API, properties, color, structure, screen reader accessibility), and render complete, accurate, and consistent spec pages directly within Figma. The system emphasizes security by running entirely locally, ensuring no proprietary design data leaves Uber's network. This significantly accelerates the documentation process and improves maintainability across multiple implementation stacks.

2025

Adding Determinism and Safety to Uber IAM Policy Changes

9/18/2025

This post introduces a Policy Simulator tool designed to enhance the safety and determinism of IAM policy changes. It addresses the high stakes of policy modifications by enabling authors to preview the impact of their proposed changes in real-time, preventing accidental outages. The post details the architecture and two-stage process of the simulator, which involves fetching access logs and replaying them on reference and proposed policy sets to identify differences in access control results. It also highlights optimizations made to the log ingestion pipeline for low-latency fetching.