Import the remade decks on stable ids and recommend one card per job, so a saved draft still loads after a chip is renamed.

The real-group profiles stay as a test, so a change that moves a pick fails.

Co-authored-by: Cursor <cursoragent@cursor.com>

data/create/customRule/_marksToConfirm.json

@@ -0,0 +1,21 @@
+{
+  "about": "Proposed and not applied. Mark changes the engine run suggested for the marks sign-off. M2 left this file on 2 November 2026. It was decided on 30 October (D7) and is now in the conflict facets and _jobs.json, with Regular debriefs first in the prevention job. M1 stays here unapplied, since it is not applied at launch (D6, 30 October). engine/results/worked-examples.md shows the picks with and without it.",
+  "fixes": [
+    {
+      "id": "M1",
+      "deck": "membership",
+      "card": "owners-and-staff",
+      "set": {
+        "orgType": {
+          "workersCoop": true
+        }
+      },
+      "traits": {
+        "fits": [
+          "nonvoters"
+        ]
+      },
+      "why": "The card says worker-owners are the owners in a worker co-op, but only for-profits are marked, so under one main kind no co-op gets it. The review's worked examples gave it to co-ops whose workers do not all own. This mark does that only if the pay answer also comes off the card, because its help line tells co-ops to tick it for worker-owners on payroll, so this revisits one cell of the 24 Oct table. It then moves 40 of the 47 co-ops in the profiles, 26 because some who work there do not vote and 14 only because they also ticked for-profit, which breaks the tie. Without it, co-ops keep Equal membership and get A voice for people without a vote beside it when some who work there do not vote."
+    }
+  ]
+}