Fix duplicate gem icon during AI streaming response

Skip rendering the empty placeholder assistant message while isLoading
is true — the GeminiTypingIndicator already shows the gem icon with
bouncing dots. Once tokens arrive and content is non-empty, the message
renders normally via GeminiMessage.

Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

frontend/components/gemini/GeminiChat.tsx

@@ -352,13 +352,19 @@ export function GeminiChat() {
             <div className="max-w-3xl mx-auto px-4 pt-8 pb-4">
               {/* 对话中也显示工单摘要 */}
               {ticketSystemConnected && <TicketSummary tickets={tickets} />}
-              {messages.map((msg) => (
+              {messages.map((msg) => {
+                // Hide empty assistant message while streaming — the typing indicator covers it
+                if (msg.role === "assistant" && msg.content === "" && isLoading) {
+                  return null;
+                }
+                return (
                   <GeminiMessage
                     key={msg.id}
                     message={msg}
                     onRegenerate={msg.role === "assistant" ? handleRegenerate : undefined}
                   />
-              ))}
+                );
+              })}
               {isLoading && <GeminiTypingIndicator />}
               <div ref={messagesEndRef} />
             </div>