Python Deserialization
Theory
Python's built-in serialization modules, such as pickle and cPickle, PyYaml, are commonly used for serializing and deserializing data. However, if the deserialization process is not properly secured, it can be exploited by attackers to execute arbitrary code or perform other malicious activities.
Practice
PyYaml Deserialization
Since PyYaml version 5.4, the default loader for load
has been switched to SafeLoader
mitigating the risks against Remote Code Execution.
The vulnerable sinks are now:
On PyYaml versions >= 5.1 (and inferior to 5.4) we can use following functions
In order for load()
and load_all()
to deserialize custom class objects, subprocess have to be imported if we use Popen in our payload. Serialized object of os.system won't works.
You can still use !!python/object/new:str
payload.
On PyYaml versions inferior to 5.1 we can use following functions
Ruamel.yaml Deserialization
To deserialize in ruamel.yaml , following methods are vulnerable to arbitrary code execution:
Pickle/cPickle Deserialization
The python pickle
and cPickle
(implementation of Pickle in C) modules, that serializes and deserializes a Python object, are vulnerables to remote code execution. If the website uses this modules, we may be able to execute arbitrary code.
With Pickle deserialization ,the following code is vulnerable to arbitrary code execution using the pickle.load()
function without proper sanitization of the input.
Jsonpickle Deserialization
Jsonpickle is a python library for serializing any arbitrary object graph into JSON.
With jsonPickle deserialization ,the following code is vulnerable to arbitrary code execution using the jsonpickle.decode()
function without proper sanitization of the input.
References
Last updated