This work presents the Governance & Integrity Model (GIM), a deterministic governance framework for behavioral security systems. GIM defines explicit authority boundaries, integrity invariants, and fail-closed operational semantics designed to ensure institutional trust, traceability, and accountability under failure and adversarial conditions. The model deliberately excludes machine learning and autonomous decision-making from core governance functions, treating them only as external advisory inputs. The paper includes a real-world case study (Melona Sentinel) demonstrating a GIM-compliant behavioral security system operating without ML, DPI, or adaptive autonomy. This record includes:- The main paper (Viewer PDF)- A supplementary note describing figures and page references for clarity without modifying the original document.