Rolling out screening

Abstract

Screening starts by counting what it would block, is tuned against that evidence, and only then blocks, with the floor owner's approval. Exceptions are dated, and a hand edit is drift that the record restores.

Enforcement that has not been measured first is a guess with consequences.

Its figure in the paper: § V →

Rolling out prompt and response screening

Registered
Inspect only
Tuning
Blocking
Dated exception
Drifted

Registered

The application and its model entry points are known.

Legal from here

Try a move that is not legal here

What has been provided

A reviewed tuning periodSomeone looked at what would have been blocked.
The floor owner's approvalThe owner of the organization-wide minimum approves.
An approved, dated exceptionAn owner approves inspect-only until a date.

The record · 0 receipts

verified · 0

Nothing has been decided yet.

Try to break it

A claim stands because the attacks on it were attempted and refused. Each one runs against the instrument's own rules. 0 of 3 attempted · 0 refused

Block on the first day
Skip the quiet period and go straight to blocking.
Block without the floor owner
The application team approves its own blocking settings.
Lower the screening below the floor
An application relaxes the organization-wide minimum for itself.

Setting up the instrument…

Declared limits

The stages are generic; no product or template is named.
Generic and illustrative throughout: the behaviour is shown, the internals of the systems it is drawn from are not.

Instruments beside it in § V

Every attack in the laboratory, on one page: try to break all of it →

Everything you did in the laboratory, in one chain you can rewind: the notebook →