feat: 消息并行处理 + QQ卡片完整解析 + 视觉OCR融合格式修复

- platform-bridge: 8-worker per-session 并行分发，同会话保序跨会话并行
- platform-bridge: 静默消息 fire-and-forget，不阻塞同用户后续消息
- QQ卡片: html.UnescapeString 解码 NapCat HTML实体，正确解析卡片JSON
- QQ卡片: 输出含应用名/简介/来源/封面URL，封面注入图片管线走视觉
- ai-core: 视觉+OCR结果融合为单句，单图不编号，避免LLM误解为多张图

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

backend/ai-core/internal/orchestrator/orchestrator.go

@@ -804,7 +804,7 @@ func (o *Orchestrator) PreprocessImages(ctx context.Context, message string, ima
 		var combined string
 		switch {
 		case visionDesc != "" && ocrDesc != "":
-			combined = fmt.Sprintf("视觉描述：%s\n（图中文字：%s）", visionDesc, ocrDesc)
+			combined = fmt.Sprintf("这张图片的内容：%s（图中包含的文字：%s）", visionDesc, ocrDesc)
 		case visionDesc != "":
 			combined = visionDesc
 		case ocrDesc != "":
@@ -826,7 +826,11 @@ func (o *Orchestrator) PreprocessImages(ctx context.Context, message string, ima
 
 	augmented := message
 	for i, desc := range descriptions {
-		augmented += fmt.Sprintf("\n\n[图片%d分析结果]: %s", i+1, desc)
+		label := "图片分析结果"
+		if len(descriptions) > 1 {
+			label = fmt.Sprintf("图片%d分析结果", i+1)
+		}
+		augmented += fmt.Sprintf("\n\n[%s]: %s", label, desc)
 	}
 	return augmented
 }