Skip to content
P‑AIProject‑AI
Systems · TS4-IIMP-001

TS4-IIMP-001

An adopted standard under Thirsty's Codex / Standards V4+ defining an evidence-governed pathway AI systems must follow before assigning emotional, motive, or safety significance to intense communication.

What it is

The governing question is: "What meaning is the user trying to keep the system from changing?" The standard requires symmetric verification — user meaning must not be overwritten without evidence, and system drift must not be fabricated or conceded without evidence — while concrete safety content still gets evaluated. It ships Python and TypeScript reference engines and an executable conformance suite.

Version history

standard v1.1.0 · repository package 1.1.0-repository.1

The 1.1.0 normative standard shipped 2026-08-04 with a nine-stage deliberation pathway and a six-criterion conformance model. A same-day repository-package release (1.1.0-repository.1) replaced a syntax-only CI check with real Draft 2020-12 JSON Schema validation and a live conformance badge.

Open CHANGELOG source

Status, limitations, and proof trail

Verified Version history and CI-hardening changes, directly checkable against the pinned CHANGELOG.

Documented Standard's normative content and reference implementations; this portal did not independently execute the conformance suite.

README proof sourcePinned repository revisionRepository