fix(workflow): Duplicate - JSON-encode JSONB fields

**Error:**
```
psycopg2.ProgrammingError: can't adapt type 'dict'
```

**Root Cause:**
- duplicate_prompt passed Python dicts directly to SQL INSERT
- JSONB fields from r2d() are already deserialized by psycopg2
- PostgreSQL expects JSON strings for JSONB columns

**Fix:**
- Added json.dumps() for all JSONB fields before INSERT:
  - stages, output_schema, question_augmentations, graph_data
- Same pattern as import function

Files changed:
- backend/routers/prompts.py: JSON-encode JSONB in duplicate_prompt

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

backend/routers/prompts.py

@@ -364,6 +364,12 @@ def duplicate_prompt(prompt_id: str, session: dict=Depends(require_admin)):
 
         new_display_name = f"{original.get('display_name') or original['name']} (Kopie)"
 
+        # Prepare JSONB fields (convert dict/list to JSON string if needed)
+        stages_json = json.dumps(original['stages']) if original.get('stages') else None
+        output_schema_json = json.dumps(original['output_schema']) if original.get('output_schema') else None
+        question_aug_json = json.dumps(original['question_augmentations']) if original.get('question_augmentations') else None
+        graph_data_json = json.dumps(original['graph_data']) if original.get('graph_data') else None
+
         cur.execute(
             """INSERT INTO ai_prompts (
                 id, name, slug, display_name, description, template, category,
@@ -379,9 +385,9 @@ def duplicate_prompt(prompt_id: str, session: dict=Depends(require_admin)):
             (new_id, new_name, new_slug, new_display_name,
              original.get('description'), original.get('template'),
              original.get('category', 'ganzheitlich'),
-             original.get('type', 'pipeline'), original.get('stages'),
-             original.get('output_format', 'text'), original.get('output_schema'),
-             original.get('question_augmentations'), original.get('graph_data'),
+             original.get('type', 'pipeline'), stages_json,
+             original.get('output_format', 'text'), output_schema_json,
+             question_aug_json, graph_data_json,
              original.get('active', True), original.get('sort_order', 0))
         )