feat(ai-core): real vision image analysis

- llm: Adapter.Vision sends multimodal chat requests (text + image_url,
  data URLs supported) and logs the call as purpose 'vision'
- gateway /images/analyze now accepts image_url or image_base64 and
  returns real analysis instead of a placeholder message
- main wires the default LLM adapter as the vision analyzer
- httptest coverage for the vision request/response shape

backend/services/gateway/api/handlers_more.go

@@ -260,16 +260,35 @@ func (r *Router) sessionExportHandler(w http.ResponseWriter, req *http.Request)
 
 func (r *Router) imageAnalyzeHandler(w http.ResponseWriter, req *http.Request) {
 	if req.Method != "POST" { writeJSON(w, 405, M{"error": "POST required"}); return }
+	if r.deps.Vision == nil { writeJSON(w, 503, M{"error": "vision not available"}); return }
 	var body struct {
-		ImageURL string `json:"image_url"`
-		Prompt   string `json:"prompt"`
+		ImageURL    string `json:"image_url"`
+		ImageBase64 string `json:"image_base64"` // raw base64 or data URL
+		Prompt      string `json:"prompt"`
 	}
 	json.NewDecoder(req.Body).Decode(&body)
-	// Stub: requires multimodal LLM (GPT-4o / Qwen-VL)
-	writeJSON(w, 200, M{
-		"analysis": "图片分析需要多模态大模型。当前为预留接口。请配置支持 vision 的 LLM provider。",
-		"image_url": body.ImageURL, "prompt": body.Prompt,
-	})
+
+	imageRef := body.ImageURL
+	if imageRef == "" && body.ImageBase64 != "" {
+		imageRef = body.ImageBase64
+		if !strings.HasPrefix(imageRef, "data:") {
+			imageRef = "data:image/jpeg;base64," + imageRef
+		}
+	}
+	if imageRef == "" {
+		writeJSON(w, 400, M{"error": "missing image_url or image_base64"})
+		return
+	}
+	if body.Prompt == "" {
+		body.Prompt = "描述这张图片的内容。"
+	}
+
+	analysis, err := r.deps.Vision.AnalyzeImage(req.Context(), body.Prompt, imageRef)
+	if err != nil {
+		writeJSON(w, 500, M{"error": err.Error()})
+		return
+	}
+	writeJSON(w, 200, M{"analysis": analysis, "prompt": body.Prompt})
 }
 
 func (r *Router) hostExecHandler(w http.ResponseWriter, req *http.Request) {

backend/services/gateway/api/router.go

@@ -91,6 +91,11 @@ type AutomationEngine interface {
 	Trigger(ctx context.Context, id string) error
 }
 
+// VisionAnalyzer analyzes images with a multimodal model.
+type VisionAnalyzer interface {
+	AnalyzeImage(ctx context.Context, prompt, imageRef string) (string, error)
+}
+
 // ChatStorer persists chat sessions and messages.
 type ChatStorer interface {
 	CreateSession(ctx context.Context, id, userID, title string) (*store.Session, error)
@@ -117,6 +122,7 @@ type Dependencies struct {
 	Automation   *AutomationStore
 	AutomationEng AutomationEngine
 	Reminders    *ReminderStore
+	Vision       VisionAnalyzer
 	PluginList   []map[string]string
 }