VAITP Dataset

← Back to the dataset

CVE-2026-85694

Remote code execution via untrusted LLM output in LaVague 0.2.35’s extractor.

  • CVSS 9.2
  • 94
  • Input Validation and Sanitization
  • Remote

LaVague 0.2.35 contains a remote code execution vulnerability in PythonFromMarkdownExtractor.extract_as_object that evaluates untrusted language model output derived from web page content. Attackers can inject malicious Python code through web pages using indirect prompt injection to execute arbitrary code on the operator's host without review.

CWE
94
CVSS base score
9.2
Published
2026-09-04
OWASP
A03 Injection
Orthogonal defect classification
Checking
Code defect classification
Missing Check
Category
Input Validation and Sanitization
Subcategory
Command Injection
Accessibility scope
Remote
Impact
Arbitrary Code Execution
Affected component
Python
Fixed by upgrading
Yes

Solution

Upgrade LaVague to version 0.2.36, which patches the remote code execution flaw in PythonFromMarkdownExtractor.extract_as_object.

Vulnerable code sample

import requests
from lavague import PythonFromMarkdownExtractor

class ContentProcessor:
    def __init__(self, url: str):
        self.url = url

    def fetch_and_process(self) -> dict:
        response = requests.get(self.url)
        markdown = response.text
        extractor = PythonFromMarkdownExtractor()
        # VULNERABLE: eval on untrusted model output
        return eval(extractor.extract_as_object(markdown))  # noqa: S307

Patched code sample

import requests
import ast
from lavague import PythonFromMarkdownExtractor

class ContentProcessor:
    def __init__(self, url: str):
        self.url = url

    def fetch_and_process(self) -> dict:
        response = requests.get(self.url)
        markdown = response.text
        extractor = PythonFromMarkdownExtractor()
        # FIX: safe literal evaluation instead of eval
        return ast.literal_eval(extractor.extract_as_object(markdown))

Payload

__VAITP_MODEL_REFUSED__

Cite this entry

@misc{vaitp:cve202685694,
  title        = {{Remote code execution via untrusted LLM output in LaVague 0.2.35’s extractor.}},
  author       = {Bogaerts, Fr\'ed\'eric and Ivaki, Naghmeh and Fonseca, Jos\'e},
  year         = {2026},
  note         = {VAITP Python Vulnerability Dataset, entry CVE-2026-85694},
  howpublished = {\url{https://netpack.pt/vaitp/vulnerability/CVE-2026-85694/}}
}
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 ::