CVE-2026-109311

vLLM before 0.29.0 Heap Overflow DoS via IAMF Audio Transcription Upload

Published Oct 12, 2026 CVSS 4.0 · 7.1 High EPSS < 1% KEV · Not listed
Description
vLLM before 0.29.0 passes uploaded audio to the PyAV/FFmpeg av.open() fallback with only a file-size check, exposing FFmpeg's IAMF parser heap overflow. Authenticated remote attackers can repeatedly upload malformed IAMF containers to /v1/audio/transcriptions, corrupting the heap and crashing the API process.

Analysis

AI-generated

Technical impact

This vulnerability occurs when the model receives uploaded audio files. The system passes the audio to FFmpeg’s IAMF parser with only a file‑size check, allowing a malformed IAMF container to trigger a heap overflow in FFmpeg. When exploited, the overflow corrupts the heap, crashes the process, and results in a denial‑of‑service. The weakness is an input validation flaw (CWE‑20) and a heap‑based buffer overflow (CWE‑787).

Affected Systems

vLLM (vllm‑project), all releases before 0.29.0.

Risk and Exploitability

The vulnerability has a CVSS score of 7.1, indicating a high impact if exploited. The EPSS score of < 1% indicates a low likelihood of exploitation in the wild, but because the flaw requires authenticated remote uploads, an attacker must have API credentials to trigger it. If credentials are compromised or shared, repeated uploads of malformed containers can repeatedly crash the service. The CVE is not listed in the CISA KEV catalog, so there is no known public exploit at the time of this analysis. Nonetheless, the nature of the overflow and the lack of input validation make it a high‑priority target for malicious actors to cause service interruption.

Remediation

No solution or workaround provided in the CVE record.

OpenCVE Recommended Actions

  • Upgrade vLLM to version 0.29.0 or later, where the IAMF parser input validation is fixed.
  • If an upgrade cannot be performed immediately, limit the exposure of the /v1/audio/transcriptions endpoint to trusted users only, require strong authentication and apply rate limiting.
  • Consider temporarily disabling audio transcription uploads or implementing a whitelist of accepted container formats to reduce the attack surface until a patch is applied.

Generated by OpenCVE AI on October 19, 2026 at 01:52 UTC.

Tracking

Sign in to view the affected projects.

Advisories

No advisories available
History
Type Values Removed Values Added
Metrics ssvc

{'options': {'Automatable': 'no', 'Exploitation': 'poc', 'Technical Impact': 'partial'}, 'version': '2.0.3'}

Type Values Removed Values Added
Description vLLM before 0.29.0 passes uploaded audio to the PyAV/FFmpeg av.open() fallback with only a file-size check, exposing FFmpeg's IAMF parser heap overflow. Authenticated remote attackers can repeatedly upload malformed IAMF containers to /v1/audio/transcriptions, corrupting the heap and crashing the API process.
Title vLLM before 0.29.0 Heap Overflow DoS via IAMF Audio Transcription Upload
First Time appeared Vllm
Vllm vllm
Weaknesses CWE-20
CPEs cpe:2.3:a:vllm:vllm:*:*:*:*:*:*:*:*
Vendors & Products Vllm
Vllm vllm
References
Metrics cvssV3_1

{'score': 6.5, 'vector': 'CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H'}

cvssV4_0

{'score': 7.1, 'vector': 'CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N'}

Weaknesses

Sources

Status
PUBLISHED
Assigner
VulnCheck
Published
Updated
Oct 18, 2026, 11:51 AM
Reserved
Oct 12, 2026, 2:14 PM
Status
Awaiting Analysis
Published
Oct 12, 2026, 5:17 PM
Modified
Oct 18, 2026, 12:17 PM
Severity
Moderate
Public date
Oct 12, 2026, 4:05 PM