补齐任务列表状态筛选和邮件HTML安全字段

server/src/main/java/cn/nianxx/thhotel/platform/message/common/result/SourceMessageConversationMessageResult.java

@@ -16,7 +16,9 @@ import java.util.List;
  * @param sourceSentAt 邮件来源发送时间，UTC
  * @param textBody 完整纯文本正文
  * @param htmlBody 完整 HTML 正文，前端展示前必须 sanitize
+ * @param htmlBodySanitized 后端第一版清洗后的 HTML，前端生产展示应优先使用
  * @param htmlSanitizeRequired 是否要求前端 sanitize HTML
+ * @param htmlRenderMode HTML 渲染建议模式，例如 SANITIZED_HTML 或 TEXT_ONLY
  * @param inlineImages 内联图片外链
  * @param attachments 附件外链
  * @param relatedOrders 关联订单摘要
@@ -39,8 +41,12 @@ public record SourceMessageConversationMessageResult(
         String textBody,
         @JsonProperty("html_body")
         String htmlBody,
+        @JsonProperty("html_body_sanitized")
+        String htmlBodySanitized,
         @JsonProperty("html_sanitize_required")
         Boolean htmlSanitizeRequired,
+        @JsonProperty("html_render_mode")
+        String htmlRenderMode,
         @JsonProperty("inline_images")
         List<SourceMessageOriginalMediaResponse> inlineImages,
         List<SourceMessageOriginalMediaResponse> attachments,

server/src/main/java/cn/nianxx/thhotel/platform/message/service/impl/SourceMessageConversationServiceImpl.java

@@ -18,6 +18,7 @@ import java.time.LocalDateTime;
 import java.time.ZoneOffset;
 import java.util.List;
 import java.util.Optional;
+import java.util.regex.Pattern;
 import org.springframework.stereotype.Service;
 import org.springframework.transaction.annotation.Transactional;
 
@@ -31,6 +32,20 @@ public class SourceMessageConversationServiceImpl implements SourceMessageConver
     private static final String ACCESS_SCENE = "source-message-conversation";
     private static final String MEDIA_TYPE_INLINE_IMAGE = "INLINE_IMAGE";
     private static final String MEDIA_TYPE_ATTACHMENT = "ATTACHMENT";
+    private static final String HTML_RENDER_MODE_SANITIZED = "SANITIZED_HTML";
+    private static final String HTML_RENDER_MODE_TEXT_ONLY = "TEXT_ONLY";
+    private static final Pattern DANGEROUS_TAG_PATTERN = Pattern.compile(
+            "(?is)<\\s*(script|style|iframe|object|embed|meta|link|base|form)[^>]*>.*?<\\s*/\\s*\\1\\s*>"
+                    + "|<\\s*(script|style|iframe|object|embed|meta|link|base|form)[^>]*/?\\s*>");
+    private static final Pattern EVENT_ATTRIBUTE_PATTERN = Pattern.compile(
+            "(?i)\\s+on[a-z0-9_-]+\\s*=\\s*(\"[^\"]*\"|'[^']*'|[^\\s>]+)");
+    private static final Pattern STYLE_ATTRIBUTE_PATTERN = Pattern.compile(
+            "(?i)\\s+style\\s*=\\s*(\"[^\"]*\"|'[^']*'|[^\\s>]+)");
+    private static final Pattern DANGEROUS_URL_ATTRIBUTE_PATTERN = Pattern.compile(
+            "(?i)\\s+(href|src|xlink:href|formaction)\\s*=\\s*"
+                    + "(\"\\s*(?:javascript|data|vbscript):[^\"]*\""
+                    + "|'\\s*(?:javascript|data|vbscript):[^']*'"
+                    + "|\\s*(?:javascript|data|vbscript):[^\\s>]+)");
 
     private final SourceMessageInboxRepository inboxRepository;
     private final List<SourceMessageRelatedContextProvider> relatedContextProviders;
@@ -101,6 +116,7 @@ public class SourceMessageConversationServiceImpl implements SourceMessageConver
     private SourceMessageConversationMessageResult toConversationMessage(SourceMessageInboxSnapshot message) {
         SourceMessageOriginalContent originalContent = readOriginalAndAudit(message);
         SourceMessageRelatedContextResult relatedContext = findRelatedContext(message);
+        String htmlBody = originalContent.htmlBody();
         return new SourceMessageConversationMessageResult(
                 message.id().toString(),
                 message.externalMessageId(),
@@ -110,8 +126,10 @@ public class SourceMessageConversationServiceImpl implements SourceMessageConver
                 UtcTimeFormatter.toUtcOffsetDateTime(message.receivedAt()),
                 UtcTimeFormatter.toUtcOffsetDateTime(message.sourceSentAt()),
                 originalContent.textBody(),
-                originalContent.htmlBody(),
+                htmlBody,
+                sanitizeHtml(htmlBody),
                 true,
+                htmlRenderMode(htmlBody),
                 filterMedia(originalContent.mediaItems(), MEDIA_TYPE_INLINE_IMAGE),
                 filterMedia(originalContent.mediaItems(), MEDIA_TYPE_ATTACHMENT),
                 relatedContext.relatedOrders(),
@@ -173,4 +191,27 @@ public class SourceMessageConversationServiceImpl implements SourceMessageConver
                 item.externalUrl(),
                 item.externalMediaId());
     }
+
+    /**
+     * 第一版 HTML 清洗：保留邮件正文基本结构，移除脚本标签、事件属性和危险协议链接。
+     */
+    private String sanitizeHtml(String htmlBody) {
+        if (htmlBody == null) {
+            return null;
+        }
+        String sanitized = DANGEROUS_TAG_PATTERN.matcher(htmlBody).replaceAll("");
+        sanitized = EVENT_ATTRIBUTE_PATTERN.matcher(sanitized).replaceAll("");
+        sanitized = STYLE_ATTRIBUTE_PATTERN.matcher(sanitized).replaceAll("");
+        return DANGEROUS_URL_ATTRIBUTE_PATTERN.matcher(sanitized).replaceAll("");
+    }
+
+    /**
+     * 返回前端渲染建议；有 HTML 时优先使用后端清洗后的安全 HTML。
+     */
+    private String htmlRenderMode(String htmlBody) {
+        if (htmlBody == null || htmlBody.isBlank()) {
+            return HTML_RENDER_MODE_TEXT_ONLY;
+        }
+        return HTML_RENDER_MODE_SANITIZED;
+    }
 }