OCF Doctrine

Glossary of Concepts

This glossary establishes the mandatory linguistic and doctrinal framework of Predictive Cyberlab. It ensures the integrity of human vulnerability measurements.

OCF Doctrinal Preeminence

The doctrine is hardcoded into the technical infrastructure (Policy-as-Code). No partner can redefine the scope of an evaluation, guaranteeing cross-company comparability.

The Cognitive Arc

The Cognitive Arc relies on a strict separation of mental tasks ("cognitive compartmentalization") to diagnose the source of a failure.

Niveau 1 (N1) : Savoir

Eliminate blind spots and make the risk mentally exist. (Pure recognition of objects/signals).

Niveau 2 (N2) : Comprendre

Produce an intelligible and structured reading of a situation. (Interpretation and correlation).

Niveau 3 (N3) : Agir

Observe the trajectory of action under constraint. Decision and commitment. (Inaction is treated as a decision).

Cognitive Blindness

State of an individual unable to recognize a risk signal (N1 Failure), rendering any subsequent analysis random.

N1 Complexity Degrees

Scale of perceptual load (D1 to D5), ranging from isolated recognition (D1) to lexical discrimination under load (D5).

Architecture & AI

Guarantees deterministic evaluation and absolute sovereignty of processed data.

Zero Data Retention (ZDR)

Data processing exclusively in RAM. No sensitive data is permanently stored. Strict "No-Training Policy".

Policy-as-Code

Embedding doctrine in master tables, making the system auditable and independent of application code.

Dual-Core Architecture

Separation between the AI Coach (immersion) and the AI Judge (deterministic evaluation), avoiding conflicts of interest.

RAG Amnésique

Volatile injection of internal procedures (PDF) for evaluation. Entirely purged at session closure.

Corridor HITL (Human-in-the-Loop)

Mandatory human oversight triggered by the AI Judge in case of a "Doubt Zone" (e.g., 65-75% score).

Steering & Metrology

Indicators linking cyber-maturity to the financial stakes of the company.

H-ROSI (Human Return on Security Investment)

Financial translation of human vulnerability reduction. Goal: Justify investments to the CFO.

ICO (Indices de Coût)

Calibration of the financial impact of simulated incidents by criticality (P1, P2, P3).

Shadow Risk

Financial exposure carried by non-evaluated populations. Drives the "Homogeneity Bonus" (> 85% coverage).

Blast Radius

Algorithmic calculation of risk propagation (N-1, N-2 dependency graph) in case of compromise.

Prescription Verticale

Capacity of a network head to impose campaigns and secure the extended Supply Chain.

HVS (Human Vulnerability Score)

Operational resilience score deterministically aggregating the Cognitive Arc.

Human Dynamics

Analysis of friction zones between cognition and action.

Interstices

Analysis of failure zones between the levels of the cognitive arc.

Inhibition

The participant KNOWS (Valid N1) but DOES NOT ACT (Failing N3).

Aveuglement

The participant DOES NOT SEE the risk (N1 Failure), making the rest obsolete.

Cohérence

Perfect alignment between understanding (N2) and engaged action (N3).

Tensions Décisionnelles

Injected conflicts forcing trade-offs between technical risk and exposure (e.g., media pressure, authority).

Dette Invisible

Vulnerability generated by a decision made under tension that creates a latent risk.

Illusion de Maîtrise

Cognitive bias where one overestimates their control capacity (tested via fake confirmations).

Co-design Levels

To preserve the SaaS model and comparability, collaboration is segmented.

1. Niveau Fermé (Non négociable)

Measurement model, H-ROSI calculation, scoring algorithms, ZDR protocol, Cognitive Arc rules.

2. Niveau Semi-ouvert (Encadré)

SSO connectors, ICO calibration, regulatory risk mapping (NIS2), RAG injection.

3. Niveau Ouvert (Libre)