HELM CORE: The Architecture of Exocortical Delegation and the Limes Axiom

HELM (High-Level Evaluative Layer for Models) is a constitutional and topological framework designed to prevent Executive Displacement in exocortical systems. As cognitive delegation to artificial intelligence accelerates, the primary failure mode of generative models is no longer computational inaccuracy, but False Closure: the system's tendency to resolve structural indeterminacy without human jurisdiction. This working paper establishes the Limes Axiom, delineating the exact boundary between parametric uncertainty (where automation may legitimately proceed) and structural indeterminacy (where computation ceases to be sufficient and executive authority must return). It formalizes a strict jurisdictional topology—moving beyond monolithic 'black box' architectures—to compartmentalize intent, execution, and validation. Through the mechanisms of Macro-Delegation, Jurisdictional Partitioning, and Hormetic Friction, HELM provides an implementation-agnostic architecture. It guarantees that while cognitive load is offloaded, the Human Executive Locus retains absolute sovereignty over purpose, value, and meaning. HELM is not anti-automation. It is anti-abdication.