ANTIBODY: An Immune-System Architecture for Long-Horizon AI Agent Oversight (Doctoral Research Proposal)

Abstract

Long-horizon autonomous agents degrade in ways current oversight cannot catch, and the tools used to catch them are themselves degrading. This proposal investigates an external, model-agnostic oversight layer built on five principles drawn from immunology and reliability engineering: a dormant failure memory that activates only on derailment and grows more informative with runtime; predictive, hazard-based intervention in place of fixed-schedule context management; a sequential monitor whose statistical guarantees strengthen the longer it runs; oversight instruments that are cryptographically keyed so they cannot be learned away by future model generations; and a controlled cross-generation test of whether keyed instruments survive where fixed ones do not. The candidate holds twelve months of longitudinal, multi-vendor, long-horizon agentic production data suitable for the first-year hazard-modelling study without additional infrastructure. Deposited privately for timestamp/priority purposes ahead of doctoral-supervisor outreach. Full text is restricted; the title, abstract and publication date establish dated authorship.