A vulnerability was discovered in the PyYAML library in versions before 5.4, where it is susceptible to arbitrary code execution when it processes untrusted YAML files through the full_load method or with the FullLoader loader. Applications that use the library to process untrusted input may be vulnerable to this flaw. This flaw allows an attacker to execute arbitrary code on the system by abusing the python/object/new constructor. This flaw is due to an incomplete fix for CVE-2020-1747.

Recent assessments:

zeroSteiner at October 29, 2020 1:51pm UTC reported:

Overview

The PyYAML library contains a vulnerability where an attacker with control over data which is loaded using the load function can trigger arbitrary Python code execution. The latest version as of this writing (5.3.1) is affected. PyYAML version 4.1 changed the default load to use safe_load by default, however this was reverted in the 4.2 release. Later, in version 5.1 the PyYAML project implemented a new default FullLoader class which prevented function calls but is in sufficient to prevent code execution.

Details

The latest versions of PyYAML (5.1+) use a default loader implementation that allow deserialization of objects and references but not functions. The builtin map is a class in Python3 and can be used to apply the builtin eval function to indirectly trigger a function call.

!!python/object/new:tuple 
- !!python/object/new:map 
  - !!python/name:eval
  - [ "RCE_HERE" ]

An attacker looking to leverage this vulnerability would need to identify that YAML is loaded by the application and then submit their crafted data to be loaded.

Solution

The PyYAML library contains a two primary alternatives to load data safely. The first is to use the load_safe alternative, the second is to specify the SafeLoader class as the Loader keyword argument to the standard load function. This is detailed in the PyYAML-Deprecation) Wiki page. Developers using PyYAML should use one of these safe methods and ensure that they are not loading data from untrusted sources.

See: <https://blog.arxenix.dev/pyyaml-cve/&gt; for more information.

Assessed Attacker Value: 3
Assessed Attacker Value: 3Assessed Attacker Value: 4