Deduplicate pair caption cast descriptors

docs/prompt-architecture-improvement-plan.md

@@ -33,6 +33,9 @@ The map audit currently sees:
   caption outputs can be debugged by category, action/position family, selected
   pair side, scene profile, position keys, and POV labels instead of only
   proving that a metadata branch was used.
+- Insta/OF side-target training captions no longer prepend shared cast
+  descriptors when the selected side row already emits its own cast prose, and
+  route simulation flags repeated cast descriptors.
 
 ## Architectural Finding
 

docs/prompt-pool-routing-map.md

@@ -1004,6 +1004,7 @@ issues for:
   action/position family, selected pair side, scene profile, position keys, and
   POV labels;
 - raw builder labels leaking into Krea output;
+- repeated cast descriptors in training-caption formatter output;
 - duplicate negative-prompt comma items;
 - softcore prompt noise;
 - POV routes emitting third-person camera text or losing first-person wording;