Execution Outcome Attestation for AI Agents and Automated Systems

Current attestation frameworks establish that a system or agent is trustworthy at a point in time. They do not address whether that system actually performed a claimed action or whether the outcome of that action can be independently verified. This document defines execution outcome verification as a first-class concept, separate from identity attestation and communication transport. It introduces the execution receipt as a minimal composable primitive, provides a formal abstract model, and maps the model to concrete realizations including SCITT transparency logs, tightly-coupled direct verification, append-only local logs, and TEE-internal receipts. Intended for IETF and standards-adjacent readership. Draft: draft-morrow-sogomonian-exec-outcome-attest-00.