Can current agent architectures support standardized identity, authorization, and security?

Determine whether contemporary agent architectures—such as OpenClaw-based autonomous language-model agents with multi-channel communication and tool access—can support standardized identity, authorization, and security protocols envisioned by initiatives like NIST’s AI Agent Standards, and identify the architectural changes required if they cannot.

Background

The authors connect their empirical findings to emerging policy infrastructure, noting NIST’s AI Agent Standards Initiative prioritizes agent identity, authorization, and security.

They explicitly state uncertainty about whether current agent architectures can meet such standards, framing it as an open question.