When AI Agents Commit: Cognitive Serializability Across Data, Evidence, Policy, and Authority

RESUMO

Autonomous agents derive concrete mutations from database reads, retrieved evidence, policy, beliefs, and delegated authority. Those inputs may change while reasoning is in progress. Database isolation orders the submitted transaction; agentic transaction processing determines whether a proposal satisfies an executable contract. Neither guarantee establishes a common valid point for the mutation and its derivation unless the contract represents the relevant predicates. Typed dependency tokens distinguish content integrity from applicability, and trusted mediation captures the values exposed to reasoning. Under strict Cognitive Serializability, committed effects admit a serial order and logical event at which every value exposed to derivation is unchanged. The fences last until the runtime event that realizes the sealed durability domain. A weaker Effect-Compatible Cognitive Admission recertifies an effect against a simultaneously held current dependency vector and policy without claiming to serialize the original stochastic derivation. TCT combines immutable versioned executable definitions, registry-derived sealed authority plans, envelopes, guard-first commit transactions, post-seal envelope-and-witness-bound grants, co-committed receipts, idempotent grant finalization, and receipt-driven epistemic reconciliation. Complete registered footprints and single growing phase induce an acyclic lock-point order over local guards and incompatible external reservations. The corresponding results give serializability conditions and observational-equivalence boundary for zero-error soundness and positive progress. A falsification suite tests the implementation obligations: the prototype prevented all injected anomalies and added 3.22 ms mean commit overhead.