diff --git a/js/modules/link_previews.js b/js/modules/link_previews.js
index 8837f1d40f..b606a42332 100644
--- a/js/modules/link_previews.js
+++ b/js/modules/link_previews.js
@@ -216,9 +216,17 @@ const ASCII_PATTERN = new RegExp('[\\u0000-\\u007F]', 'g');
 
 function isLinkSneaky(link) {
   const domain = getDomain(link);
+  // If the domain is falsy, something fishy is going on
+  if (!domain) {
+    return true;
+  }
 
-  // This is necesary because getDomain returns domains in punycode form. We check whether
-  //   it's available for the StyleGuide.
+  // Domains cannot contain encoded characters
+  if (domain.includes('%')) {
+    return true;
+  }
+
+  // This is necesary because getDomain returns domains in punycode form.
   const unicodeDomain = nodeUrl.domainToUnicode
     ? nodeUrl.domainToUnicode(domain)
     : domain;
diff --git a/test/modules/link_previews_test.js b/test/modules/link_previews_test.js
index 6054f8b5f7..78835edd95 100644
--- a/test/modules/link_previews_test.js
+++ b/test/modules/link_previews_test.js
@@ -371,5 +371,25 @@ describe('Link previews', () => {
       const actual = isLinkSneaky(link);
       assert.strictEqual(actual, true);
     });
+
+    it('returns true for =', () => {
+      const link = 'r.id=s.id';
+      assert.strictEqual(isLinkSneaky(link), true);
+    });
+
+    it('returns true for $', () => {
+      const link = 'r.id$s.id';
+      assert.strictEqual(isLinkSneaky(link), true);
+    });
+
+    it('returns true for +', () => {
+      const link = 'r.id+s.id';
+      assert.strictEqual(isLinkSneaky(link), true);
+    });
+
+    it('returns true for ^', () => {
+      const link = 'r.id^s.id';
+      assert.strictEqual(isLinkSneaky(link), true);
+    });
   });
 });