diff --git a/openverifiablellm/verify.py b/openverifiablellm/verify.py
index 0431c41..28b1076 100644
--- a/openverifiablellm/verify.py
+++ b/openverifiablellm/verify.py
@@ -331,7 +331,6 @@ def verify_preprocessing(
         logger.info("Re-running preprocessing in temp dir: %s", tmp_dir)
 
         try:
-
             env = os.environ.copy()
             env["PYTHONPATH"] = os.pathsep.join(p for p in sys.path if p)
 
@@ -410,9 +409,7 @@ def verify_preprocessing(
                 actual=reproduced_manifest.get("preprocessing_version"),
                 detail="Preprocessing version tag",
             )
-            # verify chunk size recorded in the manifest matches whatever
-            # the preprocessing run produced. this needs the reproduced
-            # manifest, so we only perform it once the file has been loaded.
+
             if "chunk_size_bytes" in manifest:
                 _check_field(
                     report, "manifest_chunk_size_bytes",
@@ -420,6 +417,12 @@ def verify_preprocessing(
                     actual=reproduced_manifest.get("chunk_size_bytes"),
                     detail="Merkle chunk size used during preprocessing",
                 )
+            else:
+                report.add(CheckResult(
+                    name="manifest_chunk_size_bytes",
+                    status=CheckStatus.SKIP,
+                    detail="Field absent from manifest (older version)",
+                ))
         else:
             report.add(CheckResult(
                 name="manifest_regenerated",
@@ -478,4 +481,4 @@ def main(argv=None):
 
 
 if __name__ == "__main__":
-    main()
+    main()
\ No newline at end of file
diff --git a/tests/test_util.py b/tests/test_util.py
index 4dcb989..430e1be 100644
--- a/tests/test_util.py
+++ b/tests/test_util.py
@@ -269,4 +269,27 @@ def test_export_and_load_merkle_proof(tmp_path):
         proof_file_path=proof_file,
         chunk_data=chunk,
         expected_root=root,
-    )
\ No newline at end of file
+    )
+
+def test_extract_text_from_xml_malformed_xml(tmp_path, monkeypatch):
+    import defusedxml.ElementTree as ET
+
+    malformed_xml_content = """<?xml version="1.0"?>
+    <mediawiki>
+      <page>
+        <revision>
+          <text>Hello [[Malformed]]
+        </revision>
+      </page>
+    </mediawiki>
+    """
+
+    input_file = tmp_path / "simplewiki-20260201-pages-malformed.xml"
+
+    with open(input_file, "w", encoding="utf-8") as f:
+        f.write(malformed_xml_content)
+
+    monkeypatch.chdir(tmp_path)
+
+    with pytest.raises(ET.ParseError):
+        utils.extract_text_from_xml(input_file)
diff --git a/tests/test_verify.py b/tests/test_verify.py
index 784c60d..1d0832c 100644
--- a/tests/test_verify.py
+++ b/tests/test_verify.py
@@ -377,12 +377,11 @@ def setUp(self):
 
     def test_merkle_checks_are_skipped(self):
         r = verify_preprocessing(self.dump, project_root=self.tmp)
-        for name in ("raw_merkle_root", "processed_merkle_root"):
+        for name in ("raw_merkle_root", "processed_merkle_root", "manifest_chunk_size_bytes"):
             c = next((x for x in r.checks if x.name == name), None)
             self.assertIsNotNone(c, f"check '{name}' not found")
             self.assertEqual(c.status, CheckStatus.SKIP)
-        # legacy manifests should not even contain the chunk-size
-        self.assertFalse(any(c.name == "manifest_chunk_size_bytes" for c in r.checks))
+            self.assertIn("Field absent from manifest (older version)", c.detail)
 
     def test_other_checks_still_pass(self):
         r = verify_preprocessing(self.dump, project_root=self.tmp)
@@ -391,4 +390,4 @@ def test_other_checks_still_pass(self):
 
 
 if __name__ == "__main__":
-    unittest.main(verbosity=2)
+    unittest.main(verbosity=2)
\ No newline at end of file