Iniciar sesión
Deutsch English Español Polski
Media

vLLM: Unauthenticated audio decompression-bomb DoS in /v1/chat/completions

  • Publicada: 16.09.2026
  • Actualizado: 11.10.2026 15:18

Summary

The audio decode-duration guard (max_duration_s, env VLLM_MAX_AUDIO_DECODE_DURATION_S, default 600s) that protects against audio decompression-bomb DoS is wired into only the speech-to-text path (/v1/audio/transcriptions). The chat audio path (/v1/chat/completions, input_audio content parts) calls the same decoder with no limit, so an unauthenticated client can submit a few-KB compressed audio file that expands to multiple GB of float32 PCM at decode time, OOM-killing the worker. This is a distinct sibling of CVE-2026-5497 (video frame-count bomb, VideoMediaIO.load_base64) and GHSA-pq5c-rjhq-qp7p (image) in the same media subsystem.

Verified against main at HEAD d78650c (2026-06-16); applicable to the latest release v0.23.0.

Details

The guard rejects long audio during decode (before allocation), implemented in vllm/multimodal/media/audio.py:

  • load_audio_pyav — metadata reject (~82-98) and live sample-count reject (~129-136)
  • load_audio_soundfile — frames reject (~165-174)

All are gated on if max_duration_s is not None.

It is passed in exactly one place — the transcription serving layer:

python
# .../speech_to_text/base/serving.py:~170-174
load_audio(buf, sr=..., max_duration_s=self.max_audio_decode_duration_s)
#   self.max_audio_decode_duration_s = envs.VLLM_MAX_AUDIO_DECODE_DURATION_S  (default 600)

The chat path never threads it:

python
# vllm/multimodal/media/audio.py:237-238
def load_bytes(self, data: bytes) -> tuple[npt.NDArray, float]:
    return load_audio(BytesIO(data), sr=None)   # no max_duration_s -> every guard above is skipped

Unauthenticated reachability chain (chat): parse_input_audio (chat_utils.py) -> parse_audio -> connector.fetch_audio -> AudioMediaIO._load_data_url -> load_base64 -> load_bytes -> load_audio(..., sr=None). The connector never passes max_duration_s, and inline data: URLs need no HTTP fetch (so VLLM_AUDIO_FETCH_TIMEOUT does not bound them). The OpenAI-compatible server has no auth by default (auth only when --api-key / VLLM_API_KEY is set).

Impact

Unauthenticated remote denial of service (availability) via memory amplification on a default-no-auth endpoint, on any deployment serving an audio-capable model. Same class and impact as the sibling CVE-2026-5497 (video). CWE-770 / CWE-409.

Fix

A fix was introduced in this MR: https://github.com/vllm-project/vllm/pull/45908

Software afectado

Haz clic en una tecnología para ver todas sus vulnerabilidades.

PlataformaPaqueteVersiones vulnerablesCorregida en
Python (PyPI) vllm < 0.24.0 0.24.0

Referencias

Puntuación CVSS 6.5/10
Vector CVSS
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H
Debilidad (CWE)
CWE-770
Plataforma
Python (PyPI)
Paquete
vllm
Publicada
16.09.2026
Actualizado
11.10.2026 15:18

Fuentes y licencias

OSV / GitHub Advisory Database - PYSEC-2026-4179 - PyPA Advisory Database; CC BY 4.0

OSV / GitHub Advisory Database - GHSA-hcwq-8wjf-3gcr - CC BY 4.0

Historial de vulnerabilidades: vllm

Todas las vulnerabilidades publicadas de este paquete (Python (PyPI)), de la más reciente a la más antigua. Muestra si es un fallo aislado o un paquete que hay que vigilar.

Todas (94) en el archivo
Entradas en total
94
Primera → última
09.2024 10.2026
Sin corrección
5
Explotada en ataques
0

Distribución por gravedad

  • Crítica 9
  • Alta 23
  • Media 47
  • Baja 4
  • Sin evaluar 11

Divulgaciones por año

Gravedad:
  1. A flaw has been found in vllm-project vLLM up to 0.26.0. This vulnerability affects unknown code of the file rust/src/parser/src/unified/gemma4.rs of the component Gemma4UnifiedParser. Executing a man
    Alta · 7.5 CVE-2026-103241 < 0.27.0 0.27.0
  2. vLLM versions 0.22.0 through 0.23.0 fail to validate stop_token_ids against vocabulary bounds in Rust HTTP and gRPC frontends, allowing out-of-vocabulary token IDs to reach MinTokensLogitsProcessor. A
    Alta · 7.5 CVE-2026-100652 >= 0.22.0, < 0.24.0 0.24.0
  3. vLLM: Unauthenticated audio decompression-bomb DoS in /v1/chat/completions esta entrada
    Media · 6.5 CVE-2026-57173 < 0.24.0 0.24.0
  4. vLLM has Remote DoS via Invalid Recovered Token Reinjection
    Alta · 7.5 CVE-2026-54234 >= 0.17.1, < 0.24.0 0.24.0
  5. vLLM: ReDoS via structured_outputs.regex compiled without timeout in xgrammar and outlines backends
    Alta · 7.5 CVE-2026-55574 < 0.24.0 0.24.0
  6. vLLM denial of service via prompt embeds on M-RoPE models
    Alta CVE-2026-55514 >= 0.12.0, < 0.24.0 0.24.0
  7. vLLM is an inference and serving engine for large language models (LLMs). Prior to 0.22.1, the vLLM Dockerfile is vulnerable to a dependency confusion attack through the flashinfer-jit-cache package.
    Alta · 8.8 CVE-2026-54232 < 0.22.1 0.22.1
  8. vLLM: OpenAI auth bypass
    Crítica · 9.1 CVE-2026-48746 >= 0.3.0, < 0.22.0 0.22.0
  9. vLLM: Security Check Bypass via assert Statement in Activation Function Loading Allows Arbitrary Code Execution
    Alta · 7.5 CVE-2026-41523 < 0.22.0 0.22.0
  10. vLLM versions 0.8.0 and later are vulnerable to an Out-of-Memory (OOM) Denial of Service (DoS) attack due to unbounded frame count processing in the `VideoMediaIO.load_base64()` method. When processin
    Alta · 7.5 CVE-2026-5497 >= 0.8.0, < 0.19.0 0.19.0
  11. vLLM has Hardcoded Trust Override in Model Files Enables RCE Despite Explicit User Opt-Out
    Alta · 8.8 CVE-2026-27893 >= 0.10.1, < 0.18.0 0.18.0
  12. vLLM has RCE In Video Processing
    Crítica · 9.8 CVE-2026-22778 >= 0.8.3, < 0.14.1 0.14.1
  13. vLLM vulnerable to Server-Side Request Forgery (SSRF) through MediaConnector
    Alta · 7.1 CVE-2026-24779 < 0.14.1 0.14.1
  14. vLLM affected by RCE via auto_map dynamic module loading during model initialization
    Alta · 8.8 CVE-2026-22807 >= 0.10.1, < 0.14.0 0.14.0
  15. vLLM introduced enhanced protection for CVE-2025-62164
    Alta · 8.8 CVE-2026-56340 >= 0.10.2, < 0.13.0 0.13.0
  16. vLLM vulnerable to remote code execution via transformers_utils/get_config
    Alta · 7.1 CVE-2025-66448 < 0.11.1 0.11.1
  17. vLLM vulnerable to DoS with incorrect shape of multimodal embedding inputs
    Alta · 6.5 CVE-2025-62372 >= 0.5.5, < 0.11.1 0.11.1
  18. vLLM deserialization vulnerability leading to DoS and potential RCE
    Alta · 8.8 CVE-2025-62164 >= 0.10.2, < 0.11.1 0.11.1
  19. vLLM is vulnerable to Server-Side Request Forgery (SSRF) through `MediaConnector` class
    Alta · 7.1 CVE-2025-6242 >= 0.5.0, < 0.11.0 0.11.0
  20. vLLM is vulnerable to timing attack at bearer auth
    Alta · 7.5 CVE-2025-59425 < 0.11.0 0.11.0
  21. vLLM has remote code execution vulnerability in the tool call parser for Qwen3-Coder
    Alta · 8.8 CVE-2025-9141 >= 0.10.0, < 0.10.1.1 0.10.1.1
  22. vllm API endpoints vulnerable to Denial of Service Attacks
    Alta · 7.5 CVE-2025-48956 >= 0.1.0, < 0.10.1.1 0.10.1.1
  23. vLLM Allows Remote Code Execution via PyNcclPipe Communication Service
    Crítica · 9.8 CVE-2025-47277 >= 0.6.5, < 0.8.5 0.8.5
  24. Remote Code Execution Vulnerability in vLLM Multi-Node Cluster Configuration
    Alta · 8.0 CVE-2025-30165 >= 0.5.2, < 0.10.0 0.10.0
  25. vLLM Vulnerable to Remote Code Execution via Mooncake Integration
    Crítica · 10.0 CVE-2025-32444 >= 0.6.5, < 0.8.5 0.8.5
  26. Data exposure via ZeroMQ on multi-node vLLM deployment
    Alta · 7.5 CVE-2025-30202 >= 0.5.2, < 0.8.5 0.8.5
  27. CVE-2025-24357 Malicious model remote code execution fix bypass with PyTorch < 2.6.0
    Crítica · 9.8 ghsa-ggpf-24jw-3fcw < 0.8.0 0.8.0
  28. vLLM deserialization vulnerability in vllm.distributed.GroupCoordinator.recv_object
    Crítica · 9.8 CVE-2024-9052 <= 0.8.1 sin corrección
  29. vLLM Deserialization of Untrusted Data vulnerability
    Crítica · 9.8 CVE-2024-11041 <= 0.6.2 sin corrección
  30. vLLM allows Remote Code Execution by Pickle Deserialization via AsyncEngineRPCServer() RPC server entrypoints
    Crítica · 9.8 CVE-2024-9053 <= 0.6.0 sin corrección
  31. vLLM Allows Remote Code Execution via Mooncake Integration
    Crítica · 9.0 CVE-2025-29783 >= 0.6.5, < 0.8.0 0.8.0
  32. vllm: Malicious model to RCE by torch.load in hf_model_weights_iterator
    Alta · 7.5 CVE-2025-24357 < 0.7.0 0.7.0
  33. vLLM denial of service vulnerability
    Alta · 7.5 CVE-2024-8768 < 0.5.5 0.5.5

Identifica las vulnerabilidades de tu aplicación

Con la integración de DockRay y el inventario de software activado, DockRay compara las versiones notificadas de los componentes con las vulnerabilidades conocidas y avisa de las coincidencias confirmadas. La cobertura depende de la tecnología compatible y de los avisos disponibles.

Conectar tu aplicación
Newsletter DockRay

La seguridad de tus aplicaciones en tu correo

Vulnerabilidades importantes, recomendaciones concretas y consejos prácticos para tus aplicaciones. Elige tus tecnologías y sigue lo que importa para tu proyecto.

  • Un resumen diario de nuevas vulnerabilidades, también las críticas.
  • Avisos sobre actualizaciones importantes.
  • Consejos prácticos sobre seguridad y monitorización de aplicaciones.

Déjalo vacío para recibir avisos de todas las tecnologías.

Política de privacidad

Suscripción gratuita. Date de baja con un clic en cualquier correo.