VAITP Dataset

← Back to the dataset

CVE-2026-105757

vLLM <0.30.0 allows structured-output bugs to crash the shared engine.

  • CVSS 6.5
  • CWE-20
  • Input Validation and Sanitization
  • Remote

vLLM is an inference and serving engine for large language models. Prior to 0.30.0, structured-output request failures can escape request-scoped validation and reach the EngineCore fatal-error path. A per-request backend mismatch can re-raise a grammar compilation exception, padding produced by the ngram_gpu speculative-decoding mode can pass a negative token to guidance validation, and the Rust frontend can admit empty structured-output values that the Python frontend rejects, allowing ordinary constrained-generation requests to terminate the shared engine. This issue is fixed in version 0.30.0.

CVSS base score
6.5
Published
2026-10-05
OWASP
A08 Software and Data Integrity Failures
Orthogonal defect classification
Checking
Code defect classification
Missing Check
Category
Input Validation and Sanitization
Subcategory
Insecure Parsing or Deserialization
Accessibility scope
Remote
Impact
Denial of Service (DoS)
Affected component
vLLM
Fixed by upgrading
Yes

Solution

Upgrade vLLM to version 0.30.0 or later.

Vulnerable code sample

import json

class EngineCore:
    def __init__(self):
        self.state = {}

    def _validate_structured_output(self, output):
        # VULNERABLE: does not reject empty structured-output values
        # Missing validation allows empty dict to propagate
        return True

    def process_request(self, request_json: str):
        request = json.loads(request_json)
        output = request.get("structured_output", {})
        if self._validate_structured_output(output):
            # Simulate downstream processing that expects non‑empty output
            result = self._run_generation(output)
            return result

    def _run_generation(self, output):
        # Placeholder for generation logic that crashes on empty output
        if not output:
            raise RuntimeError("Generated empty structured output")
        return {"generated": "text", "metadata": output}

Patched code sample

import json

class EngineCore:
    def __init__(self):
        self.state = {}

    def _validate_structured_output(self, output):
        # FIX: reject empty structured-output values before processing
        if not isinstance(output, dict) or not output:
            raise ValueError("structured_output must be a non‑empty dict")
        return True

    def process_request(self, request_json: str):
        request = json.loads(request_json)
        output = request.get("structured_output", {})
        if self._validate_structured_output(output):
            result = self._run_generation(output)
            return result

    def _run_generation(self, output):
        # Placeholder for generation logic
        return {"generated": "text", "metadata": output}

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.

Cite this entry

@misc{vaitp:cve2026105757,
  title        = {{vLLM <0.30.0 allows structured-output bugs to crash the shared engine.}},
  author       = {Bogaerts, Fr\'ed\'eric and Ivaki, Naghmeh and Fonseca, Jos\'e},
  year         = {2026},
  note         = {VAITP Python Vulnerability Dataset, entry CVE-2026-105757},
  howpublished = {\url{https://netpack.pt/vaitp/vulnerability/CVE-2026-105757/}}
}

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.