Machine-readable protocol for immutable, causally linked decision records with validator-backed integrity.

A principal-level framework for designing AI decision systems resilient to abuse, manipulation, and adversarial behaviour.

Plumb — buyside audit framework for media buying or assessment. Three-layer scoring (Optimization, Authenticity, Pricing), NERVA decision verdicts, pre-registered falsification methodology.

NERVA — decision-integrity kernel for agentic AI. Five-axis scoring, adaptive threshold τ, One-Way Brake, auditable receipts. Live v10 cockpit + mixed-state v11.

NERVA decision-integrity kernel applied to agentic shopping — gating autonomous purchase decisions with scored verdicts and receipts.

NERVA as a portable prompt — model-agnostic decision-integrity layer. Five verdicts (COMMIT/HOLD/WAIT/CONSULT/TOXIC), adaptive threshold, receipt output. Paste once, govern any LLM decision.