docs: lead decision 78, stay on refactor until Jason calls it ready; merge plan waits; row 41 (S6) is the dogfood milestone and gets review priority

Co-Authored-By: Claude Opus 5.5 <[email protected]>

docs/plans/2026-09-26_lead-decisions.md

@@ -1595,3 +1595,25 @@ which stay with him. Each item names who decided it and what happened.
       this checkout, so it picks up S6 on its next restart. Sage restarts
       it after S6 lands and records the restart. The recorded PM→coder run
       stays on a scratch data root.
+
+78. **Stay on refactor; the next milestone is a meta-harness Jason can
+    dogfood (2026-10-09).** Jason, about 23:25Z, relayed verbatim by Mos
+    (thread 38bc4c81): "Then we stay on refactor until it is ready to take
+    over." He also confirmed the stack's next milestone is a meta-harness
+    he can dogfood while project and business work goes on.
+    - refactor stays the dev branch. The refactor→next merge plan
+      (docs/plans/2026-10-08_refactor-to-next-merge-plan.md) waits as a
+      whole, steps 1 to 3 included. It starts when Jason says refactor is
+      ready. Mos is asking him what "ready" means. This replaces the
+      2026-10-13 date in today's PM standup.
+    - Row 41 (S6, #1523) is the meta-harness row. When it lands:
+      - `mosaic bus start <business> --pm` runs the PM without a window;
+      - `mosaic talk` reaches the PM from Jason's terminal;
+      - the PM launches role sessions within the business file's limits;
+      - `mosaic launches off` and `mosaic stop` work.
+      Darkwing reviews #1523 ahead of rows 13 and 40 when it arrives.
+    - Row 40 (S5, WebUI) carries on. It doesn't block dogfooding from the
+      CLI, but row 42's acceptance needs it.
+    - S6's gate uses a scratch business with no tracker, so it doesn't wait
+      on the tasks.mosaicstack.dev vs tasks.woltje.com ruling. Dogfooding
+      against the live mosaic-stack business does.