RegimeGuard

This release introduces VeilSight v5, a structural intelligence framework for detecting regime transitions, irreversible system states, and time-aware risk in multivariate time-series data. VeilSight v5 unifies four components: PhaseAtlas, a regime geometry engine; RegimeGuard, a persistence-aware irreversibility detector; RES (Regime Exposure Score), a financial and underwriting-grade risk translation; and a legally archivable underwriting artifact suitable for audit, insurance, and governance workflows. Unlike predictive or optimization-based approaches, VeilSight v5 operates on structural invariants derived from regime persistence, transition risk, entropy, and temporal drift. The system explicitly avoids pointwise forecasting and instead identifies terminal commitments—states from which recovery probability collapses under sustained structural stress. These commitments are represented symbolically via the Omega Glyph, which anchors irreversibility across domains. The framework is validated on multiple NASA CMAPSS jet engine datasets (FD001–FD004), including cross-dataset calibration, remaining useful life (RUL) correlation, lead-time survival analysis, and external dataset integration. Outputs include a live registry, alerting mechanisms, standardized JSON registries, and auto-generated PDF underwriting packets. VeilSight v5 positions structural regime intelligence as a bridge between signal analysis, operational risk, and financial decision-making, enabling deployment in safety-critical, industrial, and insurance contexts. VeilSight v5 represents a transition from symbolic emergence to decision-grade structural commitment. VeilSight v5 is a structural risk assessment system, not a forecasting or optimization model. The system operates under the following principles and assumptions: Regime Persistence Over Pointwise PredictionRisk is assessed based on sustained structural states rather than instantaneous values. Short-lived excursions are explicitly discounted. Irreversibility as a Structural PropertyCertain regime transitions exhibit persistence and entropy collapse consistent with terminal degradation. These states are treated as irreversible commitments rather than probabilistic events. Dataset-Independent NormalizationAll scoring metrics are normalized using invariant transformations, enabling cross-dataset calibration without retraining or parameter tuning. Fixed Weighting and TransparencyThe Regime Exposure Score (RES) uses fixed, documented weights. No adaptive weighting or opaque optimization is applied. Time Awareness via Drift and Lead-Time AnalysisTemporal drift penalties and lead-time survival curves quantify how early structural risk emerges relative to failure, supporting proactive intervention. No Forecast GuaranteesVeilSight does not claim to predict exact failure times. Outputs are exposure classifications intended to inform underwriting, maintenance planning, and risk governance. Auditability and Archival IntegrityAll outputs are reproducible, timestamped, and suitable for legal archiving. PDF artifacts represent system state at issuance time. Version 5 introduces Tier-3 generalization and institutional readiness. New in v5: Temporal Regime Mapping Directional arrows added to phase maps Explicit irreversibility detection Cross-Dataset Contagion Metrics Transfer risk matrices between regime populations Enables portfolio-level integrity analysis Time-Aware RES (Tier-2.5) Temporal drift penalty integrated Underwriting-grade exposure scoring External Validation Expansion NASA CMAPSS jet engine datasets integrated Confirms domain-independent failure physics Automated Underwriting Artifacts Composite regime maps Timestamped PDF audit packets Live Registry Compatibility Registry schemas finalized for alerting and governance Removed / Explicitly Excluded: Outcome prediction Optimization logic Domain-specific heuristics v5 formally positions RegimeGuard as a structural integrity measurement framework, not a forecasting system. In summary, v4 demonstrated what structure exists.v5 formalizes when structure becomes consequential. veilsight_v5/│├── README.md├── LICENSE├── CITATION.cff│├── core/│ ├── phaseatlas.py # Regime geometry + invariants│ ├── regimeguard.py # Persistence & irreversibility logic│ ├── omega_glyph.py # Terminal commitment encoding│ └── temporal_drift.py # Time-aware penalty computation│├── scoring/│ ├── res_scoring_tier25.py # RES with temporal drift│ └── underwriting_bands.py│├── validation/│ ├── cmapss_to_phaseatlas.py│ ├── nasa_jet_to_phaseatlas.py│ ├── tier2_registry.py│ ├── tier3_cross_asset.py│ └── lead_time_analysis.py│├── registry/│ ├── phaseatlas_registry.json│ ├── regime_exposure_scores.json│ └── cross_dataset_calibration.csv│├── dashboard/│ └── phaseatlas_registry_dashboard.py│├── artifacts/│ ├── underwriting_packet.pdf│ ├── lead_time_survival_curves.png│ └── regime_maps.png│└── methods/ ├── methods_and_assumptions.md └── validation_appendix.md