fix: wrap spotter viz prompts in messages array for MultiLLMResearcher

Both AI and matrix story generators were passing a plain string to
make_request(), which expects a list of message dicts. Wraps prompt
as [{"role": "user", "content": prompt}] in both generators.

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

chat_interface.py

@@ -2303,7 +2303,7 @@ To change settings, use:
             self.log_feedback(f"🎬 Generating AI Spotter Viz story ({llm_model})...")
             provider_name, model_name_str = map_llm_display_to_provider(llm_model)
             researcher = MultiLLMResearcher(provider=provider_name, model=model_name_str)
-            return researcher.make_request(prompt, max_tokens=2000, temperature=0.7)
+            return researcher.make_request([{"role": "user", "content": prompt}], max_tokens=2000, temperature=0.7)
         except Exception as e:
             self.log_feedback(f"⚠️ AI Spotter Viz story generation failed: {e}")
             return f"*(Generation failed: {e})*"
@@ -2371,7 +2371,7 @@ To change settings, use:
             self.log_feedback(f"🎬 Generating Matrix Spotter Viz story ({llm_model})...")
             provider_name, model_name_str = map_llm_display_to_provider(llm_model)
             researcher = MultiLLMResearcher(provider=provider_name, model=model_name_str)
-            return researcher.make_request(prompt, max_tokens=2000, temperature=0.6)
+            return researcher.make_request([{"role": "user", "content": prompt}], max_tokens=2000, temperature=0.6)
         except Exception as e:
             self.log_feedback(f"⚠️ Matrix Spotter Viz story generation failed: {e}")
             return f"*(Generation failed: {e})*"