Duologs Observatory — Standard / Framework
IETF working group drafts an AI Identity Management System (AIMS) for agent interactions
The IETF WIMSE Working Group published an Internet-Draft proposing a framework for applying established identity and authorisation mechanisms to AI-agent interactions. It covers agent identifiers, credentials, authentication, authorisation, policy, observability and remediation, with sections on agent mission, OAuth delegation, transaction tokens, human-in-the-loop and tool-to-service access.
Duologs Lens
AIMS is broader and more identity-oriented than Duologs. Duologs remains focused on the narrower organisational-authority problem: whether an identified machine actor, acting for an accountable principal under a defined mandate, has authority for this exact consequential action now. The IETF has not endorsed Duologs.
Board-level question
Could this happen inside your stack?