Lucene search
+L

23 matches found

osv_pysec
osv_pysec
•added 2026/10/01 4:38 p.m.•32 views

PYSEC-2026-4178 vLLM: Request-selected PyNvVideoCodec GPU decode bypasses static VRAM reservation

Summary Current vLLM main lets an inference request choose the PyNvVideoCodec GPU video decoder through mediaiokwargs.video.videobackend, but engine GPU memory reservation is computed only from static startup configuration and VLLMVIDEOLOADERBACKEND. If the server starts with the default...

6.5CVSS5.9AI score0.00583EPSS
SaveExploits1References9
PyPA
PyPA
•added 2026/10/01 4:38 p.m.•22 views

vLLM: Request-selected PyNvVideoCodec GPU decode bypasses static VRAM reservation

SummaryCurrent vLLM main lets an inference request choose the PyNvVideoCodec GPU video decoder through mediaiokwargs.video.videobackend, but engine GPU memory reservation is computed only from static startup configuration and VLLMVIDEOLOADERBACKEND. If the server starts with the default...

6.5CVSS5.9AI score0.00583EPSS
SaveExploits1References9Affected Software1
Positive Technologies
Positive Technologies
•added 2026/10/01 12:00 a.m.•11 views

PT-2026-105921

Summary Current vLLM main lets an inference request choose the PyNvVideoCodec GPU video decoder through media io kwargs.video.video backend, but engine GPU memory reservation is computed only from static startup configuration and VLLM VIDEO LOADER BACKEND. If the server starts with the default...

6.5CVSS5.9AI score0.00583EPSS
SaveExploits1References10
NVD
NVD
•added 2026/09/26 2:16 p.m.•19 views

CVE-2026-100649

vLLM before 0.29.0 contains a resource-limit bypass vulnerability in PyNvVideoCodec decoder allocation where sampler subclass shadowing allows independent counter increments. Unauthenticated attackers can select different sampler subclasses in video requests to exceed configured decoder limits an...

6.3CVSS0.00341EPSS
SaveExploits0References2
EUVD
EUVD
•added 2026/09/26 1:23 p.m.•19 views

EUVD-2026-87743

vLLM before 0.29.0 contains a resource-limit bypass vulnerability in PyNvVideoCodec decoder allocation where sampler subclass shadowing allows independent counter increments. Unauthenticated attackers can select different sampler subclasses in video requests to exceed configured decoder limits an...

6.3CVSS5.9AI score0.00341EPSS
SaveExploits0References2
Mend
Mend
•added 2026/09/26 1:23 p.m.•21 views

CVE-2026-100649

vLLM before 0.29.0 contains a resource-limit bypass vulnerability in PyNvVideoCodec decoder allocation where sampler subclass shadowing allows independent counter increments. Unauthenticated attackers can select different sampler subclasses in video requests to exceed configured decoder limits an...

6.3CVSS5.9AI score0.00341EPSS
SaveExploits0References4
Cvelist
Cvelist
•added 2026/09/26 1:23 p.m.•58 views

CVE-2026-100649 vLLM before 0.29.0 Resource Limit Bypass via Sampler Subclass

vLLM before 0.29.0 contains a resource-limit bypass vulnerability in PyNvVideoCodec decoder allocation where sampler subclass shadowing allows independent counter increments. Unauthenticated attackers can select different sampler subclasses in video requests to exceed configured decoder limits an...

6.3CVSS0.00341EPSS
SaveExploits0References2
osv_cve
osv_cve
•added 2026/09/26 1:23 p.m.•33 views

CVE-2026-100649 vLLM before 0.29.0 Resource Limit Bypass via Sampler Subclass

vLLM before 0.29.0 contains a resource-limit bypass vulnerability in PyNvVideoCodec decoder allocation where sampler subclass shadowing allows independent counter increments. Unauthenticated attackers can select different sampler subclasses in video requests to exceed configured decoder limits an...

6.3CVSS5.9AI score0.00341EPSS
SaveExploits0References4
CVE
CVE
•added 2026/09/26 1:23 p.m.•67 views

CVE-2026-100649

Versions of vLLM prior to 0.29.0 are vulnerable to a resource-limit bypass within the PyNvVideoCodec decoder allocation. The root cause is sampler subclass shadowing , which allows for independent counter increments rather than global tracking. An unauthenticated remote attacker can exploit this ...

6.3CVSS5.9AI score0.00341EPSS
SaveExploits0References2Affected Software1
Vulnrichment
Vulnrichment
•added 2026/09/26 1:23 p.m.•13 views

CVE-2026-100649 vLLM before 0.29.0 Resource Limit Bypass via Sampler Subclass

vLLM before 0.29.0 contains a resource-limit bypass vulnerability in PyNvVideoCodec decoder allocation where sampler subclass shadowing allows independent counter increments. Unauthenticated attackers can select different sampler subclasses in video requests to exceed configured decoder limits an...

6.3CVSS5.9AI score0.00341EPSS
SaveExploits0References2
Rapid7 Vulnerability Database (full)
Rapid7 Vulnerability Database (full)
•added 2026/09/26 12:00 a.m.•18 views

CVE-2026-100649: Allocation of Resources Without Limits or Throttling

vLLM before 0.29.0 contains a resource-limit bypass vulnerability in PyNvVideoCodec decoder allocation where sampler subclass shadowing allows independent counter increments. Unauthenticated attackers can select different sampler subclasses in video requests to exceed configured decoder limits an...

6.3CVSS5.9AI score0.00341EPSS
SaveExploits0References4
Positive Technologies
Positive Technologies
•added 2026/09/26 12:00 a.m.•20 views

PT-2026-99320

Name of the Vulnerable Software and Affected Versions vLLM versions prior to 0.29.0 Description A resource-limit bypass exists in the PyNvVideoCodec decoder allocation. This issue occurs because sampler subclass shadowing allows independent counter increments. Unauthenticated attackers can select...

6.3CVSS5.8AI score0.00341EPSS
SaveExploits0References8
Github Security Blog
Github Security Blog
•added 2026/09/17 5:17 p.m.•46 views

vLLM: Request-selected PyNvVideoCodec GPU decode bypasses static VRAM reservation

Summary Current vLLM main lets an inference request choose the PyNvVideoCodec GPU video decoder through mediaiokwargs.video.videobackend, but engine GPU memory reservation is computed only from static startup configuration and VLLMVIDEOLOADERBACKEND. If the server starts with the default...

6.5CVSS5.9AI score0.00583EPSS
SaveExploits1References7Affected Software1
EUVD
EUVD
•added 2026/09/17 5:17 p.m.•26 views

EUVD-2026-80935

vLLM: Request-selected PyNvVideoCodec GPU decode bypasses static VRAM reservation...

6.5CVSS5.9AI score0.00583EPSS
SaveExploits1References6
osv_ghsa
osv_ghsa
•added 2026/09/17 5:17 p.m.•22 views

GHSA-8PW2-6JV3-MJ5J vLLM: Request-selected PyNvVideoCodec GPU decode bypasses static VRAM reservation

Summary Current vLLM main lets an inference request choose the PyNvVideoCodec GPU video decoder through mediaiokwargs.video.videobackend, but engine GPU memory reservation is computed only from static startup configuration and VLLMVIDEOLOADERBACKEND. If the server starts with the default...

6.5CVSS5.9AI score0.00583EPSS
SaveExploits1References7
RedhatCVE
RedhatCVE
•added 2026/09/16 6:35 p.m.•26 views

CVE-2026-69147

A flaw was found in vLLM, an inference and serving engine for large language models. An attacker can exploit this by submitting specially crafted video requests that force the use of the PyNvVideoCodec GPU decoder. This bypasses the engine's static GPU memory reservation, allowing the attacker to...

6.5CVSS5.7AI score0.00583EPSS
SaveExploits1References7
NVD
NVD
•added 2026/09/16 6:17 p.m.•25 views

CVE-2026-69147

vLLM is an inference and serving engine for large language models. Prior to 0.28.0, request bodies for Chat Completions and Responses can set mediaiokwargs.video.videobackend to pynvvideocodec, and MediaConnector.fetchvideo forwards that choice to VideoMediaIO even when startup configuration...

6.5CVSS0.00583EPSS
SaveExploits1References4
osv_cve
osv_cve
•added 2026/09/16 5:49 p.m.•19 views

CVE-2026-69147 vLLM: Request-selected PyNvVideoCodec GPU decode bypasses static VRAM reservation

vLLM is an inference and serving engine for large language models. Prior to 0.28.0, request bodies for Chat Completions and Responses can set mediaiokwargs.video.videobackend to pynvvideocodec, and MediaConnector.fetchvideo forwards that choice to VideoMediaIO even when startup configuration...

6.5CVSS5.9AI score0.00583EPSS
SaveExploits1References6
Vulnrichment
Vulnrichment
•added 2026/09/16 5:49 p.m.•22 views

CVE-2026-69147 vLLM: Request-selected PyNvVideoCodec GPU decode bypasses static VRAM reservation

vLLM is an inference and serving engine for large language models. Prior to 0.28.0, request bodies for Chat Completions and Responses can set mediaiokwargs.video.videobackend to pynvvideocodec, and MediaConnector.fetchvideo forwards that choice to VideoMediaIO even when startup configuration...

6.5CVSS5.9AI score0.00583EPSS
SaveExploits1References4
Cvelist
Cvelist
•added 2026/09/16 5:49 p.m.•57 views

CVE-2026-69147 vLLM: Request-selected PyNvVideoCodec GPU decode bypasses static VRAM reservation

vLLM is an inference and serving engine for large language models. Prior to 0.28.0, request bodies for Chat Completions and Responses can set mediaiokwargs.video.videobackend to pynvvideocodec, and MediaConnector.fetchvideo forwards that choice to VideoMediaIO even when startup configuration...

6.5CVSS0.00583EPSS
SaveExploits1References4
Rows per page
Query Builder