fix(doc): GetDoc 识别完整 HTML 文档为 html 格式

- 内容以 <!DOCTYPE/html 开头且标为 markdown 时按 html 返回，避免 marked 拆坏内嵌 </style> 污染站点样式（配合前端 @layer 隔离）
- TempLate 同步 GVSite 最新构建产物

TempLate/index.html

@@ -5,7 +5,7 @@
   <meta name="viewport" content="width=device-width, initial-scale=1.0">
   <title>银河锋锐（GVSDS）</title>
   <link rel="icon" href="https://hb.cdn.gvsds.com/assets/icon/ICON.png">
-  <script type="module" crossorigin src="/GVSite/index-gTeqxgRk.js"></script>
+  <script type="module" crossorigin src="/GVSite/index-CTnafLRB.js"></script>
   <link rel="stylesheet" crossorigin href="/GVSite/index-DRWk7H-y.css">
 </head>
 <body>

backup/www/index.html

@@ -5,7 +5,7 @@
   <meta name="viewport" content="width=device-width, initial-scale=1.0">
   <title>银河锋锐（GVSDS）</title>
   <link rel="icon" href="https://hb.cdn.gvsds.com/assets/icon/ICON.png">
-  <script type="module" crossorigin src="/GVSite/index-DhQua3qJ.js"></script>
+  <script type="module" crossorigin src="/GVSite/index-DrabGyKU.js"></script>
   <link rel="stylesheet" crossorigin href="/GVSite/index-DRWk7H-y.css">
 </head>
 <body>

gvsds/apps/core/views.py

@@ -161,10 +161,15 @@ def _doc_api(file_path):
         doc = DocManager.GetDocument(uuid_bytes)
         if doc is None:
             return JsonResponse({'error': 'not_found'}, status=404)
+        content = doc.DocContent or ''
+        doc_format = doc.DocFormat
+        # 数据兜底：内容实为完整 HTML 文档时按 html 渲染（marked 会把内嵌 </style> 拆坏，污染站点样式）
+        if doc_format == 'markdown' and content.lstrip().lower().startswith(('<!doctype', '<html')):
+            doc_format = 'html'
         return JsonResponse({
             'DocTitle': doc.DocTitle or '未命名文档',
-            'DocContent': doc.DocContent or '',
-            'DocFormat': doc.DocFormat,
+            'DocContent': content,
+            'DocFormat': doc_format,
             'DocPermission': 1,
             'DocDecrypt': True,
         })