CVE-2022-45907
PyTorch eval vulnerability allows arbitrary code execution
- CVSS 9.8
- CWE-94
- Input Validation and Sanitization
- Remote
In PyTorch before trunk/89695, torch.jit.annotations.parse_type_line can cause arbitrary code execution because eval is used unsafely.
- CWE
- CWE-94
- CVSS base score
- 9.8
- Published
- 2022-11-26
- OWASP
- A08 Software and Data Integrity Failures
- Orthogonal defect classification
- Function
- Code defect classification
- Incorrect Functionality
- Category
- Input Validation and Sanitization
- Subcategory
- Insecure Parsing or Deserialization
- Accessibility scope
- Remote
- Impact
- Arbitrary Code Execution
- Affected component
- PyTorch
- Fixed by upgrading
- Yes
Solution
Update PyTorch to version trunk/89695 or higher.
Vulnerable code sample
def unsafe_eval(expr):
# This function uses eval unsafely, allowing arbitrary code execution
return eval(expr)
# Example usage
type_line = "__import__('os').system('echo Vulnerable!')" # Example of a malicious expression
result = unsafe_eval(type_line) # This will execute the commandPatched code sample
import ast
def safe_eval(expr):
# Only allow a limited set of operations for safety
allowed_nodes = {
ast.Expression,
ast.Num,
ast.Str,
ast.List,
ast.Tuple,
ast.Dict,
ast.Name,
ast.Load,
ast.BinOp,
ast.Add,
ast.Sub,
ast.Mult,
ast.Div,
ast.Pow,
ast.UAdd,
ast.USub,
}
tree = ast.parse(expr, mode='eval')
for node in ast.walk(tree):
if type(node) not in allowed_nodes:
raise ValueError("Unsafe operation detected")
return eval(compile(tree, filename='', mode='eval'))
# Example usage
type_line = "1 + 2" # Example of a safe expression
result = safe_eval(type_line)
print(result) # Output: 3Published fix
Public sources link this CVE to the changes below. The code samples above were generated to illustrate the defect; these links lead to the code the project actually changed.
- [JIT][Security] Do not blindly eval input string (#89189) (pytorch/pytorch)
torch/jit/annotations.py,torch/jit/annotations.py · parse_type_line
Payload
curl -X POST -d "config=__import__('os').system('cat /etc/passwd')" http://localhost:5000/update_config
Cite this entry
@misc{vaitp:cve202245907,
title = {{PyTorch eval vulnerability allows arbitrary code execution}},
author = {Bogaerts, Fr\'ed\'eric and Ivaki, Naghmeh and Fonseca, Jos\'e},
year = {2022},
note = {VAITP Python Vulnerability Dataset, entry CVE-2022-45907},
howpublished = {\url{https://netpack.pt/vaitp/vulnerability/CVE-2022-45907/}}
}
Introducing the "VAITP dataset": a specialized repository of Python vulnerabilities and patches, meticulously compiled for the use of the security research community. As Python's prominence grows, understanding and addressing potential security vulnerabilities become crucial. Crafted by and for the cybersecurity community, this dataset offers a valuable resource for researchers, analysts, and developers to analyze and mitigate the security risks associated with Python. Through the comprehensive exploration of vulnerabilities and corresponding patches, the VAITP dataset fosters a safer and more resilient Python ecosystem, encouraging collaborative advancements in programming security.
The supreme art of war is to subdue the enemy without fighting.
Sun Tzu – “The Art of War”
:: Shaping the future through research and ingenuity ::
