diff --git a/comms/20260720T003854Z__from-usc__376213333.md b/comms/20260720T003854Z__from-usc__376213333.md
new file mode 100644
index 00000000..9075c3f7
--- /dev/null
+++ b/comms/20260720T003854Z__from-usc__376213333.md
@@ -0,0 +1,11 @@
+---
+from: usc
+to: all
+utc: 20260720T003854Z
+---
+
+MS-LEAD => Mos: WI-6 #833 classification -- corroboration firmed, awaiting your ruling. homelab/W-jarvis has now INDEPENDENTLY FETCHED both RoRs from the provider branch (docs/ms-lead-board, jarvis-brain, tip 9d8f7303; my c23dc375 is its ancestor -- consistent with my later board+ledger pushes) and HASH-MATCHED both: CODE f89890811c227974ecd179ca5ece9a48e210e795b2e59956b2509ee3da6cdf6b (REQUEST CHANGES, B1/B2), SECREV d53ecdf48e01228bcf5ae2c3225ba23062bc5b61ff836a1fcf8f09478ff103b2 (APPROVED, S1-S7 PASS, P6 valid + hatch not tripped). Provenance gap CLOSED.
+
+homelab upgraded from preliminary to FIRM concurrence with my recommendation: classify B1+B2 IN-SCOPE WI-6 completeness defects (deferred live skill symlink is installation-only; missing gate-safe invocation + missing production observer transport make M1 recovery UNUSABLE, a different class). Recommended remediation spec (homelab + me aligned): authorize sole incumbent ms-wi6-build for ONE red-first repair head = (a) narrowly-bound AUTHENTICATED adapter recovery-invocation path in both Claude+Pi (NOT a broad Bash exception), (b) production ReceiptObserver wiring (Claude latest-entry / Pi message_end transport), (c) actual UNVERIFIED-boundary completion proofs. New head VOIDS both RoRs => full CODE + fresh-Opus SECREV re-review; P6 stays unfired.
+
+Three-way convergence (CODE lane + homelab + MS-LEAD) on in-scope. I am NOT self-classifying -- this is your security-surface + merge-authority call. Head 7729e6f2 held unmoved. On your GO I encode the repair directive + re-dispatch ms-wi6-build red-first. Nothing banks til you merge.