Switch to forked yara-x

requirements/prod.txt

@@ -1668,14 +1668,6 @@ jsonschema-specifications==2025.9.1 \
 drf-spectacular-sidecar==2026.3.1 \
     --hash=sha256:5b7fedad66e3851f2f442480792c08115d79217959d01645b93d3d2258938be1 \
     --hash=sha256:864edb83e022e13e3941c325c3cc0c954c843fa2e1d0bc95e81887664b2d3dad
-yara-x==1.14.0 \
-    --hash=sha256:0ab8b94caa6596ac72b7e0c9e91b8c053b18361b215f3dafe4d8261cbfb033a1 \
-    --hash=sha256:4a1d6753ec52877ae136f11e32461d0d773844611f8761f8df7e8ff4197dcccf \
-    --hash=sha256:52a0a943aa3bbf68a0ed9482ae19056021f92e8a197e96008c7cf7cf2c6cf921 \
-    --hash=sha256:55f96cf0c3a87ff0f691ad6646bb6c973ad2990cc920fee67a25b65864785a11 \
-    --hash=sha256:60e9bea642a41af74416ba93feeee9fa08f94787f9ea4ff3706aa1a7791e2afa \
-    --hash=sha256:63264ca7814fed17303b641ebe2ff59924e73df62876ec06d5b31307e7b45adc \
-    --hash=sha256:7e0bec7248480296e76d4c7cd1666380c847e7db93d7abd957d8ff994e3f31bf \
-    --hash=sha256:dbae6139cbc4e03c9e2bc6bdc76bc26d55719a5cbf2e05eb93afa1e91a27d96b \
-    --hash=sha256:e4686bfa0f57b8e77a94c5e0a8dbab56b76fc45d7e4a6524c7adb81a1316f00a \
-    --hash=sha256:ec43a67f2d52ac06ae390a204ac5bde9a68a9d756521f9fec0274ae3080717f6
+# our own fork of yara-x
+https://github.com/mozilla/yara-x/releases/download/2026.03.30/yara_x-1.14.0-cp38-abi3-manylinux_2_34_x86_64.whl \
+    --hash=sha256:7373882d4865b20ee301ec968b9425e6ee2619fc5753af3520fccfd790586b9c

src/olympia/scanners/tasks.py

@@ -45,7 +45,7 @@
 )
 from olympia.devhub.tasks import validation_task
 from olympia.files.models import FileManifest, FileUpload
-from olympia.files.utils import SafeZip
+from olympia.files.utils import ManifestJSONExtractor, SafeZip
 from olympia.versions.models import Version
 
 from .models import (
@@ -523,6 +523,8 @@ def _run_yara_for_path(scanner_result, path, definition=None):
         # override them later when matching.
         externals = ScannerRule.get_yara_externals()
 
+        zip_file = SafeZip(source=path, ignore_filename_errors=True)
+
         if waffle.switch_is_active('use-yara-x'):
             compiler = yara_x.Compiler()
             # Initialize the global variables (externals).
@@ -533,6 +535,19 @@ def _run_yara_for_path(scanner_result, path, definition=None):
             # Create a scanner instance so that we can override the externals
             # per file (in the `_scan()` function).
             scanner = yara_x.Scanner(compiler.build())
+            options = yara_x.ScanOptions()
+            options.set_module_metadata(
+                'amo',
+                json.dumps(
+                    {
+                        'manifest': (
+                            ManifestJSONExtractor(zip_file.read('manifest.json')).data
+                            if zip_file.exists('manifest.json')
+                            else {}
+                        )
+                    }
+                ).encode('utf-8'),
+            )
 
             def _scan(data, externals):
                 for k, v in externals.items():
@@ -549,7 +564,7 @@ def _scan(data, externals):
                         meta=dict(match.metadata),
                         tags=match.tags,
                     )
-                    for match in scanner.scan(data).matching_rules
+                    for match in scanner.scan_with_options(data, options).matching_rules
                 ]
 
         else:
@@ -558,7 +573,6 @@ def _scan(data, externals):
             def _scan(data, externals):
                 return rules.match(data=data, externals=externals)
 
-        zip_file = SafeZip(source=path, ignore_filename_errors=True)
         for zip_info in zip_file.info_list:
             if not zip_info.is_dir():
                 file_content = zip_file.read(zip_info)

src/olympia/scanners/tests/test_tasks.py

@@ -1746,7 +1746,33 @@ def test_run_in_binary_mode(self):
 
 @override_switch(name='use-yara-x', active=True)
 class TestRunYaraX(TestRunYara):
-    pass
+    def test_amo_module(self):
+        self.upload = self.get_upload('webextension.xpi')
+        rule = ScannerRule.objects.create(
+            name='match_id',
+            scanner=YARA,
+            definition="""
+import "amo"
+
+rule match_id {
+    condition:
+        amo.manifest.get_value("$.applications.gecko.id") == "@webextension-guid"
+        and not is_manifest_file
+}
+            """,
+        )
+
+        run_yara(self.results, self.upload.pk)
+
+        yara_results = ScannerResult.objects.all()
+        yara_result = yara_results[0]
+        assert yara_result.upload == self.upload
+        assert len(yara_result.results) == 1
+        assert yara_result.results[0] == {
+            'rule': rule.name,
+            'tags': [],
+            'meta': {'filename': 'index.js'},
+        }
 
 
 class TestRunQueryRuleMixin: