CVE-2021-37678

YAML deserialization in TensorFlow allows arbitrary code execution

• CVSS 8.8
• CWE-502 Deserialization of Untrusted Data
• Input Validation and Sanitization
• Local

TensorFlow is an end-to-end open source platform for machine learning. In affected versions TensorFlow and Keras can be tricked to perform arbitrary code execution when deserializing a Keras model from YAML format. The [implementation](https://github.com/tensorflow/tensorflow/blob/460e000de3a83278fb00b61a16d161b1964f15f4/tensorflow/python/keras/saving/model_config.py#L66-L104) uses `yaml.unsafe_load` which can perform arbitrary code execution on the input. Given that YAML format support requires a significant amount of work, we have removed it for now. We have patched the issue in GitHub commit 23d6383eb6c14084a8fc3bdf164043b974818012. The fix will be included in TensorFlow 2.6.0. We will also cherrypick this commit on TensorFlow 2.5.1, TensorFlow 2.4.3, and TensorFlow 2.3.4, as these are also affected and still in supported range.

CVE

CVE-2021-37678

CWE

CWE-502 Deserialization of Untrusted Data

CVSS base score

8.8

Published

2021-08-12

OWASP

A08 Software and Data Integrity Failures

Orthogonal defect classification

Timing/Serialization

Code defect classification

Serialization Issues

Category

Input Validation and Sanitization

Subcategory

Insecure Parsing or Deserialization

Accessibility scope

Local

Impact

Arbitrary Code Execution

Affected component

TensorFlow

Fixed by upgrading

Yes

Solution

Update TensorFlow to version 2.6.0 or higher.

import yaml

def load_yaml_model(yaml_string):
    """Vulnerable function that demonstrates the security issue."""
    # VULNERABLE: This code is susceptible to command injection
    return yaml.unsafe_load(yaml_string)

yaml = """
!!python/object/apply:os.system ["cat /etc/passwd"]
"""

model_config = load_yaml_model(yaml)
print(model_config)

import yaml

def load_yaml_model(yaml_string):
    """Secure function that fixes the vulnerability."""
    # SECURE: This version prevents command injection
    return yaml.safe_load(yaml_string)

yaml = """
!!python/object/apply:os.system ["cat /etc/passwd"]
"""

model_config = load_yaml_model(yaml)
print(model_config)

@misc{vaitp:cve202137678,
  title        = {{YAML deserialization in TensorFlow allows arbitrary code execution}},
  author       = {Bogaerts, Fr\'ed\'eric and Ivaki, Naghmeh and Fonseca, Jos\'e},
  year         = {2021},
  note         = {VAITP Python Vulnerability Dataset, entry CVE-2021-37678},
  howpublished = {\url{https://netpack.pt/vaitp/vulnerability/CVE-2021-37678/}}
}