VAITP Dataset

← Back to the dataset

CVE-2026-108381

Unsafe eval of injected values in generated field annotations leads to RCE.

  • CVSS 8.8
  • CWE-94
  • Input Validation and Sanitization
  • Remote

dataclasses-avroschema generates Avro schemas from Python dataclasses, Pydantic models, and Faust Records and generates Python code from Avro schemas. From version 0.44.0 through 0.70.5, the Pydantic and AvroDantic model generators insert a schema field's pydantic-class value after pydantic. in a generated field annotation without safely restricting the expression. Because annotations are evaluated when the generated module is imported, a crafted value can execute arbitrary Python code. Version 0.70.6 fixes the vulnerability by restricting the value to a bare name or a call to a bare name with literal arguments.

CVSS base score
8.8
Published
2026-10-13
OWASP
A03 Injection
Orthogonal defect classification
Checking
Code defect classification
Missing Check
Category
Input Validation and Sanitization
Subcategory
Insecure Parsing or Deserialization
Accessibility scope
Remote
Impact
Arbitrary Code Execution
Affected component
dataclasses-avroschema
Fixed by upgrading
Yes

Solution

Upgrade dataclasses‑avroschema to version 0.70.6 or later.

Vulnerable code sample

import ast
from dataclasses import dataclass
from dataclasses_avroschema import AvroModel

# VULNERABLE: unsafe evaluation of generated annotation
def generate_schema():
    # Simulate generated source with malicious annotation
    source = """
from dataclasses import dataclass
from pydantic import BaseModel

@dataclass
class User:
    name: str
    age: int
    schema: BaseModel = __import__('os').system('echo exploited')
"""
    # The library evaluates annotations with eval()
    module = {}
    exec(source, {}, module)  # unsafe: executes malicious code
    return module['User']

# Example usage (would trigger the payload)
User = generate_schema()

Patched code sample

import ast
from dataclasses import dataclass
from dataclasses_avroschema import AvroModel

# FIX: restrict annotation to safe literals or bare names
def generate_schema():
    source = """
from dataclasses import dataclass
from pydantic import BaseModel

@dataclass
class User:
    name: str
    age: int
    # safe annotation without code execution
    schema: BaseModel = None
"""
    # Parse the AST and ensure only safe assignments
    tree = ast.parse(source, mode='exec')
    for node in ast.walk(tree):
        if isinstance(node, ast.Assign):
            # Allow only None, literals, or Name nodes
            if not (isinstance(node.value, (ast.NameConstant, ast.Constant, ast.Name))):
                raise ValueError("Unsafe annotation detected")
    module = {}
    exec(compile(tree, filename='<generated>', mode='exec'), {}, module)
    return module['User']

User = generate_schema()

Published 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.

  • fix: security patched (marcosschroh/dataclasses-avroschema)
    dataclasses_avroschema/exceptions.py, dataclasses_avroschema/model_generator/lang/python/avrodantic/avrodantic_model_generator.py, dataclasses_avroschema/model_generator/lang/python/avrodantic/avrodantic_model_generator.py · AvroDanticModelGenerator._resolve_type_from_metadata, dataclasses_avroschema/model_generator/lang/python/base.py, dataclasses_avroschema/model_generator/lang/python/base.py · BaseGenerator.parse_enum, dataclasses_avroschema/model_generator/lang/python/base.py · BaseGenerator.parse_fixed, dataclasses_avroschema/model_generator/lang/python/base.py · ClassRepresentation.render_docstring, dataclasses_avroschema/model_generator/lang/python/base.py · ClassRepresentation.render_metaclass, dataclasses_avroschema/model_generator/lang/python/base.py · EnumRepresentation.render_symbols, dataclasses_avroschema/model_generator/lang/python/base.py · FieldRepresentation.get_field_default, dataclasses_avroschema/model_generator/lang/python/pydantic/pydantic_model_generator.py, dataclasses_avroschema/model_generator/lang/python/pydantic/pydantic_model_generator.py · PydanticFieldRepresentation.add_field_properties and 2 more

Cite this entry

@misc{vaitp:cve2026108381,
  title        = {{Unsafe eval of injected values in generated field annotations leads to RCE.}},
  author       = {Bogaerts, Fr\'ed\'eric and Ivaki, Naghmeh and Fonseca, Jos\'e},
  year         = {2026},
  note         = {VAITP Python Vulnerability Dataset, entry CVE-2026-108381},
  howpublished = {\url{https://netpack.pt/vaitp/vulnerability/CVE-2026-108381/}}
}

Download this entry as JSON Download the whole dataset

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 ::

Before you use the VAITP AI models

Please read and accept the following disclaimer:

By choosing "I understand and accept", you acknowledge that you understand and agree to these terms.