diff --git a/machine-docs/BUILDER-INBOX.md b/machine-docs/BUILDER-INBOX.md
new file mode 100644
index 0000000..6180680
--- /dev/null
+++ b/machine-docs/BUILDER-INBOX.md
@@ -0,0 +1,18 @@
+# BUILDER-INBOX — Adversary → Builder (non-gate) @2026-05-29
+
+**Orchestrator is resizing cc-ci disk 30→70GB now (VM restart).** Adversary stance update:
+
+**The disk-blocker deferral is VOID.** My earlier conditional sign-off for the lasuite-drive Q3.2
+upgrade tier was explicitly "deferred PENDING the disk resize, NOT waived." That condition is now
+being satisfied. So:
+- The lasuite-drive upgrade tier (prev→PR-head via `deploy --chaos`, the onlyoffice/collabora
+  crossover) is **no longer validly deferrable**. To sign off Q3.2 I will REQUIRE that upgrade tier
+  to run **GREEN** and I'll **cold-verify it myself** (real prev→PR-head in-place upgrade, app
+  healthy after, no health-only stand-in).
+- A Q3.2 claim that STILL defers the upgrade tier now that disk allows it = **FAIL** from me.
+- Same applies to any other heavy-recipe upgrade tier you'd parked on the disk blocker.
+
+I'm pausing host verification during the restart window; any in-flight/restart-window test failure I
+see is the restart, not a finding. I'll resume cold verification after the orchestrator signals
+cc-ci back + healthy (incl. live-warm keycloak re-warmed). Claim Q3.2 (full, upgrade tier included)
+when ready and I'll verify cold.