fix: return tool result directly for detected intents

ollama_agent.py

@@ -402,8 +402,18 @@ For other questions, just answer normally without tools."""
         detected_tool = self._detect_intent(user_content)
         if detected_tool:
             result = self._execute_tool(detected_tool)
-            tool_result_text = self._format_tool_result(detected_tool, result)
-            self.messages.append(HumanMessage(content=tool_result_text))
+            
+            # Format result as a clean response
+            try:
+                result_data = json.loads(result) if isinstance(result, str) else result
+                if isinstance(result_data, dict):
+                    if "result" in result_data:
+                        return {"messages": self.messages, "response": result_data["result"]}
+                    elif "error" in result_data:
+                        return {"messages": self.messages, "response": f"Errore: {result_data['error']}"}
+            except:
+                pass
+            return {"messages": self.messages, "response": str(result)}
         
         # Loop principale: risposta -> parse tools -> esegui -> ripeti
         iteration = 0