fix: preserve angle-bracketed emails (Name <a@b>) from HTML tag stripper

Teams users pasting addresses in RFC-5322 mailbox form (e.g. from Outlook)
produced messages where the bot silently extracted zero emails, because
<email@domain> was matched and removed by the generic <[^>]+> tag stripper
before the email regex ran. Tighten the stripper to skip <...> spans that
contain an @, since real HTML tags never do.

Author: rofe

src/parser.js

@@ -27,7 +27,7 @@ function decodeEntities(text) {
 function stripMarkup(text) {
   return decodeEntities(text)
     .replace(/<at[^>]*>.*?<\/at>/gi, ' ')
-    .replace(/<[^>]+>/g, ' ')
+    .replace(/<(?![^>]*@)[^>]+>/g, ' ')
     .replace(TLD_GLUE_RE, '.$1 ');
 }