← Policy gallery

ipl_cert_drift_monitored

fix-certification safety pure_temporal iplimandrafix-certificationdriftmonitoring

In production, live sessions must be checked against the IPL model; any drift between specification and running system must surface as an observable failure and be resolved or raised as an incident.

Formula

G(drift_detected → F(drift_resolved ∨ Incident))

Hover any operator or symbol for what it means.

Why it matters

Digital Twin & Drift Detection. A certification is a snapshot; systems and specs change. Overlaying the IPL model's regions onto live sessions turns spec-vs-system divergence into an observable signal — but only if every detected drift is then driven to resolution or escalated, never silently ignored.

Examples

passes drift_detected → root-caused and drift_resolved (or raised as an Incident)

fails drift_detected against the IPL model and the session continues with no response

Use it

ponens policies add ipl_cert_drift_monitored --into ./trace.json
ponens trace check ./trace.json