Prevents arbitrary code execution during python/object/new constructor

lib/yaml/constructor.py

@@ -56,6 +56,14 @@ def check_data(self):
         # If there are more documents available?
         return self.check_node()
 
+    def check_state_key(self, key):
+        """Block special attributes/methods from being set in a newly created
+        object, to prevent user-controlled methods from being called during
+        deserialization"""
+        if self.get_state_keys_blacklist_regexp().match(key):
+            raise ConstructorError(None, None,
+                "blacklisted key '%s' in instance state found" % (key,), None)
+
     def get_data(self):
         # Construct and return the next document.
         if self.check_node():
@@ -495,6 +503,16 @@ def construct_undefined(self, node):
         SafeConstructor.construct_undefined)
 
 class FullConstructor(SafeConstructor):
+    # 'extend' is blacklisted because it is used by
+    # construct_python_object_apply to add `listitems` to a newly generate
+    # python instance
+    def get_state_keys_blacklist(self):
+        return ['^extend$', '^__.*__$']
+
+    def get_state_keys_blacklist_regexp(self):
+        if not hasattr(self, 'state_keys_blacklist_regexp'):
+            self.state_keys_blacklist_regexp = re.compile('(' + '|'.join(self.get_state_keys_blacklist()) + ')')
+        return self.state_keys_blacklist_regexp
 
     def construct_python_str(self, node):
         return self.construct_scalar(node).encode('utf-8')
@@ -590,18 +608,23 @@ def make_python_instance(self, suffix, node,
         else:
             return cls(*args, **kwds)
 
-    def set_python_instance_state(self, instance, state):
+    def set_python_instance_state(self, instance, state, unsafe=False):
         if hasattr(instance, '__setstate__'):
             instance.__setstate__(state)
         else:
             slotstate = {}
             if isinstance(state, tuple) and len(state) == 2:
                 state, slotstate = state
             if hasattr(instance, '__dict__'):
+                if not unsafe and state:
+                    for key in state.keys():
+                        self.check_state_key(key)
                 instance.__dict__.update(state)
             elif state:
                 slotstate.update(state)
             for key, value in slotstate.items():
+                if not unsafe:
+                    self.check_state_key(key)
                 setattr(instance, key, value)
 
     def construct_python_object(self, suffix, node):
@@ -723,6 +746,10 @@ def make_python_instance(self, suffix, node, args=None, kwds=None, newobj=False)
         return super(UnsafeConstructor, self).make_python_instance(
             suffix, node, args, kwds, newobj, unsafe=True)
 
+    def set_python_instance_state(self, instance, state):
+        return super(UnsafeConstructor, self).set_python_instance_state(
+            instance, state, unsafe=True)
+
 UnsafeConstructor.add_multi_constructor(
     u'tag:yaml.org,2002:python/object/apply:',
     UnsafeConstructor.construct_python_object_apply)

lib3/yaml/constructor.py

@@ -31,6 +31,14 @@ def check_data(self):
         # If there are more documents available?
         return self.check_node()
 
+    def check_state_key(self, key):
+        """Block special attributes/methods from being set in a newly created
+        object, to prevent user-controlled methods from being called during
+        deserialization"""
+        if self.get_state_keys_blacklist_regexp().match(key):
+            raise ConstructorError(None, None,
+                "blacklisted key '%s' in instance state found" % (key,), None)
+
     def get_data(self):
         # Construct and return the next document.
         if self.check_node():
@@ -472,6 +480,16 @@ def construct_undefined(self, node):
         SafeConstructor.construct_undefined)
 
 class FullConstructor(SafeConstructor):
+    # 'extend' is blacklisted because it is used by
+    # construct_python_object_apply to add `listitems` to a newly generate
+    # python instance
+    def get_state_keys_blacklist(self):
+        return ['^extend$', '^__.*__$']
+
+    def get_state_keys_blacklist_regexp(self):
+        if not hasattr(self, 'state_keys_blacklist_regexp'):
+            self.state_keys_blacklist_regexp = re.compile('(' + '|'.join(self.get_state_keys_blacklist()) + ')')
+        return self.state_keys_blacklist_regexp
 
     def construct_python_str(self, node):
         return self.construct_scalar(node)
@@ -574,18 +592,23 @@ def make_python_instance(self, suffix, node,
         else:
             return cls(*args, **kwds)
 
-    def set_python_instance_state(self, instance, state):
+    def set_python_instance_state(self, instance, state, unsafe=False):
         if hasattr(instance, '__setstate__'):
             instance.__setstate__(state)
         else:
             slotstate = {}
             if isinstance(state, tuple) and len(state) == 2:
                 state, slotstate = state
             if hasattr(instance, '__dict__'):
+                if not unsafe and state:
+                    for key in state.keys():
+                        self.check_state_key(key)
                 instance.__dict__.update(state)
             elif state:
                 slotstate.update(state)
             for key, value in slotstate.items():
+                if not unsafe:
+                    self.check_state_key(key)
                 setattr(instance, key, value)
 
     def construct_python_object(self, suffix, node):
@@ -711,6 +734,10 @@ def make_python_instance(self, suffix, node, args=None, kwds=None, newobj=False)
         return super(UnsafeConstructor, self).make_python_instance(
             suffix, node, args, kwds, newobj, unsafe=True)
 
+    def set_python_instance_state(self, instance, state):
+        return super(UnsafeConstructor, self).set_python_instance_state(
+            instance, state, unsafe=True)
+
 UnsafeConstructor.add_multi_constructor(
     'tag:yaml.org,2002:python/object/apply:',
     UnsafeConstructor.construct_python_object_apply)

tests/data/myfullloader.subclass_blacklist

@@ -0,0 +1,5 @@
+- !!python/object/new:yaml.MappingNode
+  args:
+  state:
+    mymethod: test
+    wrong_method: test2

tests/data/overwrite-state-new-constructor.loader-error

@@ -0,0 +1,5 @@
+- !!python/object/new:yaml.MappingNode
+  args:
+  state:
+    extend: test
+    __test__: test

tests/lib/test_constructor.py

@@ -17,7 +17,7 @@ def _make_objects():
     global MyLoader, MyDumper, MyTestClass1, MyTestClass2, MyTestClass3, YAMLObject1, YAMLObject2,  \
             AnObject, AnInstance, AState, ACustomState, InitArgs, InitArgsWithState,    \
             NewArgs, NewArgsWithState, Reduce, ReduceWithState, Slots, MyInt, MyList, MyDict,  \
-            FixedOffset, today, execute
+            FixedOffset, today, execute, MyFullLoader
 
     class MyLoader(yaml.Loader):
         pass
@@ -235,6 +235,10 @@ def tzname(self, dt):
         def dst(self, dt):
             return datetime.timedelta(0)
 
+    class MyFullLoader(yaml.FullLoader):
+        def get_state_keys_blacklist(self):
+            return super(MyFullLoader, self).get_state_keys_blacklist() + ['^mymethod$', '^wrong_.*$']
+
     today = datetime.date.today()
 
 def _load_code(expression):
@@ -289,6 +293,18 @@ def test_constructor_types(data_filename, code_filename, verbose=False):
 
 test_constructor_types.unittest = ['.data', '.code']
 
+def test_subclass_blacklist_types(data_filename, verbose=False):
+    _make_objects()
+    try:
+        yaml.load(open(data_filename, 'rb').read(), MyFullLoader)
+    except yaml.YAMLError as exc:
+        if verbose:
+            print("%s:" % exc.__class__.__name__, exc)
+    else:
+        raise AssertionError("expected an exception")
+
+test_subclass_blacklist_types.unittest = ['.subclass_blacklist']
+
 if __name__ == '__main__':
     import sys, test_constructor
     sys.modules['test_constructor'] = sys.modules['__main__']

tests/lib3/test_constructor.py

@@ -14,7 +14,7 @@ def _make_objects():
     global MyLoader, MyDumper, MyTestClass1, MyTestClass2, MyTestClass3, YAMLObject1, YAMLObject2,  \
             AnObject, AnInstance, AState, ACustomState, InitArgs, InitArgsWithState,    \
             NewArgs, NewArgsWithState, Reduce, ReduceWithState, Slots, MyInt, MyList, MyDict,  \
-            FixedOffset, today, execute
+            FixedOffset, today, execute, MyFullLoader
 
     class MyLoader(yaml.Loader):
         pass
@@ -222,6 +222,10 @@ def tzname(self, dt):
         def dst(self, dt):
             return datetime.timedelta(0)
 
+    class MyFullLoader(yaml.FullLoader):
+        def get_state_keys_blacklist(self):
+            return super().get_state_keys_blacklist() + ['^mymethod$', '^wrong_.*$']
+
     today = datetime.date.today()
 
 def _load_code(expression):
@@ -274,6 +278,18 @@ def test_constructor_types(data_filename, code_filename, verbose=False):
 
 test_constructor_types.unittest = ['.data', '.code']
 
+def test_subclass_blacklist_types(data_filename, verbose=False):
+    _make_objects()
+    try:
+        yaml.load(open(data_filename, 'rb').read(), MyFullLoader)
+    except yaml.YAMLError as exc:
+        if verbose:
+            print("%s:" % exc.__class__.__name__, exc)
+    else:
+        raise AssertionError("expected an exception")
+
+test_subclass_blacklist_types.unittest = ['.subclass_blacklist']
+
 if __name__ == '__main__':
     import sys, test_constructor
     sys.modules['test_constructor'] = sys.modules['__main__']