feat(studio): R3 step_respond、运行页 UX 与 R4 思考流式

新增 worldbook 步骤 LLM 回复（thinking/draft/questions/evaluation），
运行页聊天区与产物/选项联动；评价区标签改为「评价与修改建议」；
流式开关开启时 NDJSON 推送思考内容，完成后拆分更新各字段。

Co-authored-by: Cursor <cursoragent@cursor.com>

data/agent/skill_templates.json

@@ -52,7 +52,22 @@
       "supportsLoopUntilSatisfied": true,
       "supportsInputs": true,
       "supportsInsertion": true,
-      "supportsScoring": true
+      "supportsScoring": true,
+      "configDefaults": {
+        "stepGoal": "",
+        "thinkingPrompt": "====== 思考流程 ======\nStep1: 简短确认任务性质（新设计/修改）\nStep2: 阅读绑定角色/世界书与上文引用\nStep3: 按步骤目标起草世界书条目\nStep4: 对照评价维度自检并优化表述\n\n（核心目的、评价标准与优化建议由系统在运行时自动注入，无需在此填写）",
+        "insertion": {
+          "position": 1,
+          "activationType": "permanent",
+          "key": "",
+          "keysecondary": "",
+          "comment": ""
+        },
+        "scoring": {
+          "enabled": true,
+          "dimensions": []
+        }
+      }
     }
   ]
 }

data/agent/templates/builtin.studio.example/pipeline.json

@@ -29,7 +29,6 @@
       "id": "aesthetic",
       "skillId": "studio.worldbook_entry",
       "displayName": "整体美学",
-      "niche": "aesthetic_tone",
       "enabled": true,
       "loopUntilSatisfied": true,
       "config": {
@@ -64,7 +63,6 @@
       "id": "persona",
       "skillId": "studio.worldbook_entry",
       "displayName": "具体人设",
-      "niche": "persona_detail",
       "enabled": true,
       "loopUntilSatisfied": true,
       "config": {