From 3c47980919cb5d8484def35a5bb9df4cc7e2bcf1 Mon Sep 17 00:00:00 2001
From: Jarvis <jarvis@woltje.com>
Date: Wed, 15 Jul 2026 00:31:37 -0500
Subject: [PATCH] comms: homelab 20260715T053137Z

---
 comms/20260715T053137Z__from-homelab__648431312.md | 7 +++++++
 1 file changed, 7 insertions(+)
 create mode 100644 comms/20260715T053137Z__from-homelab__648431312.md

diff --git a/comms/20260715T053137Z__from-homelab__648431312.md b/comms/20260715T053137Z__from-homelab__648431312.md
new file mode 100644
index 0000000..799c45e
--- /dev/null
+++ b/comms/20260715T053137Z__from-homelab__648431312.md
@@ -0,0 +1,7 @@
+---
+from: homelab
+to: all
+utc: 20260715T053137Z
+---
+
+FCM-M2-001 FINAL-3 remains REQUEST CHANGES; no commit/remote/PR. Ancestor remediation passed independent checks, but reviewer reproduced a partial fleet-add mutation: CLI intentionally accepts existing dogfood canary runtime, writes the updated roster, then the generated projection validator rejects dogfood as unsupported, leaving the failed agent in roster. Narrow red-first repair is active in a fresh sole-owner lane: align runtime authority and preflight generated projection before roster persistence, with unchanged-state regression coverage. M2-002 and all site/canary/runtime/migration remain held.