Record E2 target applicability constraints

Assistant: codex
Assistant-Model: gpt-5.6-sol
Assistant-Session: 01a0260c-4067-7052-9647-ad000d576e38

docs/target-applicability.md

@@ -0,0 +1,47 @@
+# Target applicability for WHITEHAT-WP-0001
+
+Status: active review, 2026-08-22
+
+The E2 attacker model requires an authenticated identity that is legitimately
+authorized for tenant A and not tenant B. A target name alone is insufficient;
+the target must expose a boundary where that identity exists.
+
+## audit-core — applicable
+
+`audit-core` binds each sender token to allowed sources, tenant set, read and
+write capabilities at one application choke point. Two distinct ordinary
+sender identities scoped one per fixture tenant instantiate the model without
+operator or database privilege. Its event-id, correlation and append routes
+therefore form a valid E2 differential.
+
+The target owner reviewed the route and identity design on 2026-08-21. Formal
+acknowledgement and value-safe provisioning remain required by
+`WH-ENG-20260821-AUDIT-E2`; no target traffic is permitted while its
+`target_owner_acknowledged_at` is empty.
+
+## tenant-engine — E2 applicability not established
+
+The original workplan named `tenant-engine` because its declaration said an E2
+mechanism was implemented. Cold review of the actual contract shows that this
+does not supply the attacker identity T03 assumes:
+
+- `tenancy.yaml` declares current E1/A0, not E2, and says inbound actor/tenant
+  values are not verified.
+- The flex-auth policy registers the platform-wide `tenant-engine` service
+  subject for administrative actions. It is not a tenant-A subject and is
+  deliberately allowed to address tenant B.
+- Tenant registry enumeration is the service's function and is an explicit
+  placement exception. Cross-tenant access by its authorized platform operator
+  is not a tenant-boundary failure.
+- Substituting an unregistered actor would test caller authorization (A), not
+  E2. Substituting the all-tenant service actor would give the modeled attacker
+  authority it is supposed not to hold.
+
+Therefore no production packet is sent under
+`WH-ENG-20260821-TENANT-E2` until the target owner names a real authenticated
+tenant-A identity with no tenant-B authority. If none exists, the honest
+artifact is `not applicable`, and the workplan acceptance must stop naming
+tenant-engine as though it claimed an exercisable E2 boundary. Its existing
+A/I gaps remain target-owned; whitehat must not relabel them to make this plan
+finish.
+

workplans/WHITEHAT-WP-0001-cross-tenant-evidence.md

@@ -151,6 +151,15 @@ route inventories live in `probe-packs/tenant-engine-e2.json` and
 `probe-packs/audit-core-e2.json`. The two target runs remain gated on complete
 engagement records, owner acknowledgements and executable identity adapters.
 
+Production engagements were operator-approved on 2026-08-21 and committed in
+`engagements/`. Audit-core accepted the route/identity shape and reviewed the
+adapter; its first review caught a missing idempotency header before traffic,
+which is corrected and regression-tested. `docs/target-applicability.md`
+records a separate issue: tenant-engine currently exposes no authenticated
+tenant-A identity with no tenant-B authority, so its E2 applicability is
+pending owner confirmation rather than being faked with its all-tenant service
+operator or an unauthenticated actor.
+
 ### T04 — Prove the probes fail
 
 ```task