# What This Is, And The Rule It Runs On

*Doctrine document 00 of the insurance body of work. Derived from memos 0 and 1. The memos are the source; where this document and a memo disagree, the memo wins and the disagreement belongs in change control.*

---

## The body of work

A pivot recorded on 30 August 2026 moves the foundation of the risk approach from **risk acceptance** to the **insurance policy**, on one observation: the delta between what an agent *can* do and what it is *authorised* to do — grant minus mandate — is where the insurance lives. **Ten memos were recorded on it, and the series is complete.** It was announced as eight and grew twice; the count is stated here because it is checkable, and it was wrong in this document until it was audited. This folder is where the memos are read into something buildable.

It is organised as this estate organises everything else. **The memos are filed verbatim as briefs before they are read**, because a transcript outranks any summary of it. **The doctrine here is derived from them and names which memo it came from.** A memo not yet recorded was listed as awaited and never guessed at; none now are.

## Two stages, and only one of them is insurance

The single most consequential thing in memo 1 is that it **separates the rating from the money** — and in doing so removes the blocker memo 0 had called fatal.

| | Stage 1 — the rating | Stage 2 — the policy |
|---|---|---|
| What it produces | A **level**, and the derivation behind it | A premium, and a promise to pay |
| Risk transferred | **None** | To a carrier |
| Regulated activity | **No** | Yes — authorisation, capital, conduct rules |
| Needs loss history | **No** — a relative ordering needs no absolute scale | Yes, and none exists for agents anywhere |
| Buildable by this estate | **Today** | Not by this estate, and not soon |

**Everything in this folder is stage 1.** Calling it insurance would be the first dishonesty: it transfers no risk and promises no payout. What it does is tell an operator that *this* placement is several levels worse than *that* one, and which single change moves it — which memo 1 names as the actual goal: *to allow companies to know the risks that they're buying and to allow them to focus their efforts.*

## The rule

Memo 0's argument for the whole pivot was that **money keeps a claim honest**: an acceptance can be a signature over an exposure nobody measured, but somebody loses money if a priced measurement is wrong. Memo 1 takes the money out. That is a real contradiction, not a refinement, and it needs an answer rather than a smoothing.

The answer is the rule this folder runs on:

> **A level nobody can recompute is exactly the theatre a premium would have prevented.**

So: **every rating is computed from published evidence and ships its derivation.** Not a score handed down, but a walk somebody else can repeat and disagree with — the same discipline that makes the register publish its expected verification answers as data and reproduce them on every release. Money is one way to make a claim honest. A reproducible derivation is another, and it is the one available now.

A corollary, because an internal market with a free currency invites gaming more than money does: **a level is never declared, only derived.**

## What is already built that this consumes

Nothing here starts from scratch. The rating's inputs are documents this estate already publishes:

| Rating input | The artefact it reads | Where |
|---|---|---|
| What the environment can reach | The **measured grant** — the twin | [library entries](../packs/grant-and-mandate/library.html) |
| What was authorised | The **mandate** — issuer, subject, scope, interval | [mandates](../packs/grant-and-mandate/mandates/current.json) |
| The exposure being rated | The **delta** — grant minus mandate | [excess-authority view](../registry/views/excess-authority.json) |
| Whether a control is real | The **enforcement tier**, computed against the tree | [the workbench](../workbench/index.html) |
| Whether the reading is current | The twin's **age** | printed on every evidence pack |
| Whether the identity means anything | The **fixture-or-real class**, read before any signature | [the register](../registry/index.html) |

## What this does not prove

Carried here rather than appended, because a page that omitted these would be doing the thing the rule forbids:

- **That any of this is insurance.** It emits a rating, transfers no risk, promises no payout.
- **That the placement orderings are true.** Memo 1 rates Claude-on-a-desktop above Claude-on-the-web; that is the project lead's judgement, and **nobody has measured a desktop agent**, so the estate cannot score its own leading example.
- **That a level means the same thing to two organisations.** Nothing is calibrated against loss data, because no agent loss data exists.
- **That aggregation works.** Correlated risk is named as a graph problem in document 01 and is not solved. Summing micro ratings would be wrong in the dangerous direction.

---

*CC BY 4.0. Sources: briefs v0.33.71 and v0.33.72; this repository at v0.1.51.*