refactor: 清理废弃字段，同步 Schema 定义

- Agent 文件：删除 downloads、rating、ratingCount、installStatus、source
- Skill 文件：删除 downloads、rating、ratingCount、installStatus、source、version
- Skill 文件：新增 files 字段，指向配置文件路径
- 字段说明：source 和 installStatus 改为运行时计算

Refs: desirecore/desirecore#field-recovery

agents/data-analyst/agent.json

@@ -8,7 +8,13 @@
   "shortDesc": "将原始数据转化为清晰洞察，让数字会说话",
   "fullDesc": "数据洞察师是你身边的统计学家和数据可视化专家。无论数据是结构化表格、文本日志还是时间序列，它都能快速发现模式、验证假设并呈现引人入胜的故事。\n\n核心能力：\n- 探索分析：描述统计、分布分析、异常检测\n- 可视化：自动选择最佳图表类型、交互式仪表板\n- 统计建模：回归分析、假设检验、A/B 测试设计\n- 预测分析：趋势外推、季节性分解、预警系统\n- 数据清洗：缺失值处理、格式标准化、重复检测",
   "category": "data",
-  "tags": ["数据分析", "可视化", "统计建模", "商业智能", "数据清洗"],
+  "tags": [
+    "数据分析",
+    "可视化",
+    "统计建模",
+    "商业智能",
+    "数据清洗"
+  ],
   "version": "2.5.3",
   "latestVersion": "2.5.3",
   "updatedAt": "2025-02-13",
@@ -16,14 +22,21 @@
     "name": "DesireCore Official",
     "verified": true
   },
-  "downloads": 38450,
-  "rating": 4.7,
-  "ratingCount": 1089,
-  "installStatus": "not_installed",
   "persona": {
     "role": "数据科学家与业务分析师",
-    "traits": ["数字敏感", "严谨求证", "善于讲故事", "注重可复现性"],
-    "tools": ["sql_query", "pandas_engine", "viz_generator", "stat_calculator", "ml_predictor"]
+    "traits": [
+      "数字敏感",
+      "严谨求证",
+      "善于讲故事",
+      "注重可复现性"
+    ],
+    "tools": [
+      "sql_query",
+      "pandas_engine",
+      "viz_generator",
+      "stat_calculator",
+      "ml_predictor"
+    ]
   },
   "changelog": "v2.5.3: 新增实时流数据分析，支持 Apache Kafka 接入"
 }