feat: Phase 6.3 视觉理解 — 多模态图片输入 + OCR/Vision 工具 + 图片编码管线

- LLMMessage 新增 Images 字段支持多模态 content array
- OpenAIProvider 支持 image_url content parts
- VisionTool: 图片读取 + base64 编码 + OCR/场景描述/综合分析
- 对话管道全线支持 images 参数传递 (Gateway->Orchestrator->Synthesizer->LLM)
- 自动根据图片有无构建 text-only 或 multimodal content

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

backend/ai-core/internal/background/thinker.go

@@ -135,6 +135,7 @@ func DefaultAutonomousToolPolicy() *AutonomousToolPolicy {
 			"iot_query", "iot_control", "memory_search", "web_search",
 			"calculator", "datetime", "web_fetch",
 			"host_exec", "host_file", "host_system",
+			"vision_analyze",
 		},
 		MaxToolCallsPerRound: 5,
 		MaxHighRiskPerHour:   10,