From ddae41b7c9f31f8a6c0100a3a4ee15f6c55f57f0 Mon Sep 17 00:00:00 2001
From: 0xMarcio <marc@syntax9.ai>
Date: Thu, 22 Oct 2026 21:54:31 +0200
Subject: [PATCH] Replace unreadable captures instead of reusing them

---
 scripts/archive_pocs.py    |  4 ++++
 tests/test_archive_pocs.py | 15 +++++++++++++++
 2 files changed, 19 insertions(+)

diff --git a/scripts/archive_pocs.py b/scripts/archive_pocs.py
index 7e236133a01..0e66612ce1e 100644
--- a/scripts/archive_pocs.py
+++ b/scripts/archive_pocs.py
@@ -383,6 +383,10 @@ def verify_snapshot(snap: dict, original: str, expected_blobs=(), repo_id=None,
                 return False
         snap['verified_at'] = now()
         return True
+    except ArchiveError as exc:
+        if str(exc) in {'response-too-large', 'unsupported-response-encoding', 'invalid-compressed-response'}:
+            return False
+        raise
     except error.HTTPError as exc:
         status = exc.code
         retry = exc.headers.get('Retry-After', '')
diff --git a/tests/test_archive_pocs.py b/tests/test_archive_pocs.py
index 7c656875a09..d40e517cd10 100644
--- a/tests/test_archive_pocs.py
+++ b/tests/test_archive_pocs.py
@@ -503,6 +503,21 @@ class ArchiveTests(unittest.TestCase):
                 with self.assertRaisesRegex(archive.ArchiveError, 'invalid-compressed-response'):
                     archive.response_bytes(response, 1000)
 
+    def test_corrupt_replay_is_replaced_instead_of_reusing_the_failed_capture(self):
+        snap = snapshot()
+        response = io.BytesIO(gzip.compress(b'poc')[:-8])
+        response.headers = {'Content-Encoding': 'gzip'}
+        response.status = 200
+        response.url = snap['url'].replace(snap['timestamp'] + '/', snap['timestamp'] + 'id_/')
+        opener = MagicMock()
+        opener.open.return_value = response
+        row = {**self.target, 'candidate': snap}
+        with patch.object(archive.request, 'build_opener', return_value=opener):
+            archive.verify_candidate(row)
+        self.assertTrue(row['bad_capture'])
+        self.assertNotIn('candidate', row)
+        self.assertFalse(archive.valid_snapshot(row))
+
 
 if __name__ == '__main__':
     unittest.main()