AI Agent Identity and Access
AI agents are workload identities with added agentic attributes: their identity layer is converging while authorization is not, and the consequential incidents are access failures. This report frames agent risk by type and argues for runtime authorization as the key control to manage risk.
