What happened
On 12 October 2026 (datatracker 'Last updated 2026-10-12'), a new individual Internet-Draft 'External Authorization and Bounded CLOCKOUT for Agentic Systems' (draft-watts-agent-clockout-profile-00, Deonté Watts / GoodShyt Group) was posted. It specifies a proposed operational profile that separates agent proposals from authoritative authorization, effect admission, computation termination and bounded closure, defining enforcement assumptions, revocation barriers, descendant handling, status evidence and fault/conformance tests. It explicitly complements existing authorization and revocation mechanisms and defines no new token format.
Why it matters
This is an early, concrete proposal addressing a recognized agentic-AI security gap: agents continuing to act after their authorization has changed or been revoked (stale/in-flight effects, queued jobs, delegated credentials). It contributes requirements language and test criteria to the ongoing IETF effort to standardise agent identity, authorization and audit (alongside the OAuth agent-authorization and agent-identity drafts).
Action needed
Security and identity engineers designing agentic systems should review the draft for revocation-barrier and effect-admission requirements, and track whether it moves toward WG adoption; feed back via the datatracker comment mechanism.