<?xml version="1.0" encoding="UTF-8"?>
<reference anchor="I-D.sato-soos-idp" target="https://datatracker.ietf.org/doc/html/draft-sato-soos-idp-04">
   <front>
      <title>The Intent Declaration Primitive (IDP) for Agentic AI Systems</title>
      <author initials="" surname="Sato" fullname="Tom Sato">
         <organization>MyAuberge K.K.</organization>
      </author>
      <date month="June" day="10" year="2026" />
      <abstract>
	 <t>   Every action an AI agent takes is a decision.  Right now, none of
   those decisions are signed.

   AI agents operating in automated workflows take actions without any
   normative mechanism for expressing why those actions are being taken.
   Access tokens declare what an agent is permitted to do; no existing
   standard declares what the agent believes it is doing, on what
   reasoning basis, and with what level of confidence, at the moment of
   action.  This document defines the Intent Declaration Primitive
   (IDP): a structured per-transition declaration submitted by an AI
   agent to the Governing Enforcement Component (GEC) at each action
   step of an execution loop.  The IDP is committed to a tamper-evident
   Event Log before the action executes, enabling post-hoc review of
   agent reasoning, richer authorization policy evaluation, and enriched
   denial responses that guide agent behaviour.  The IDP also provides
   the technical basis for compliance with EU AI Act Article 12 logging
   requirements for high-risk AI systems.  This document adds the IDP
   Commitment Verification mechanism, by which the GEC verifies at each
   state transition that the agent&#x27;s actual action matches its declared
   intent.  This document also adds the RETRY_CONTINUATION reasoning
   basis type and the prior_denial_count Cedar context attribute, which
   together address computational inefficiency arising from uninformed
   agent retry loops.  This document further defines, in Section 5.6,
   the schemas of the outcome Event Log entries (STATE_TRANSITIONED,
   CEDAR_DENY_RECORDED, and ACTION_RESULT_RECORDED) that record the
   result of each governed transition, completing the pre-action/post-
   action audit pair that gives the IDP its full auditability property.
   This document renames the Governing Kernel role to Governing
   Enforcement Component (GEC) to accommodate application-layer,
   isolated-process, and kernel-level implementations under a common
   conformance model.  A Conformance Levels section (Section 9) defines
   three implementation profiles (Application, Isolated, and Kernel)
   with graduated non-suppressibility guarantees.  A data_residency
   field is added to the IDP structure (Section 4.1) governing analytics
   tier eligibility.  Section 3.5 states the Event Log as Operational
   Intelligence design intent.

	 </t>
      </abstract>
   </front>
   <seriesInfo name="Internet-Draft" value="draft-sato-soos-idp-04" />
   
</reference>
