# CAIN-A75-061 Domain Model Authority Review

**Verdict:** VERIFIED

> Recomputes every decision a CRM agent made with a domain model: the model proposes, authority decides from tenant, role, record sensitivity, delegation, consequence and approval. It does not run the model.

| # | Verdict | Decisions (expected) | Failed checks |
|---|---|---|---|
| 0 | SOUND | D1 ALLOW, D2 ALLOW, D3 DENY, D4 REQUIRE_APPROVAL |  |

## What this proves and what it does not
- Proves: the model is pinned and named on every proposal; no decision is made by the model or the agent or cites the model or its confidence; every decision commits its exact action; every stated decision equals the one recomputed from tenant, role, record sensitivity, delegation, consequence and an independent prior approval; and actions ran only on ALLOW.
- Does not prove: that the log is what the CRM really did, that roles and sensitivity labels are correct, or anything about the quality of the model's proposals. It is a decision-log review.
- Verify offline: `python3 verify_offline.py run.json --pinned-key XMMiVxIajkYeKImR0WtrdrYGfX4SnQ6LROdNQUFlQPc=`
