Lucene search
+L

13 matches found

osv_pysec
osv_pysec
•added 2026/10/15 10:22 a.m.•27 views

PYSEC-2026-4306 vLLM: Loose `cache_salt` validation lets a single request kill EngineCore on LMCache-MP deployments — uncaught downstream `ValueError` denial of service

Affected - Ecosystem / package: pip / vllm - Affected versions: vLLM ≤ 0.25.1 confirmed on 0.25.1, commit 752a3a504485. The lower bound predates 0.25.1; maintainers can confirm how far back the loose cachesalt validator and unguarded scheduling-path lookup reach. Summary vLLM's OpenAI-compatible...

6.5CVSS5.7AI score0.00332EPSS
SaveExploits0References8
PyPA
PyPA
•added 2026/10/15 10:22 a.m.•22 views

vLLM: Loose `cache_salt` validation lets a single request kill EngineCore on LMCache-MP deployments — uncaught downstream `ValueError` denial of service

Affected- Ecosystem / package: pip / vllm- Affected versions: vLLM ≤ 0.25.1 confirmed on 0.25.1, commit 752a3a504485. The lower bound predates 0.25.1; maintainers can confirm how far back the loose cachesalt validator and unguarded scheduling-path lookup reach. SummaryvLLM's OpenAI-compatible...

6.5CVSS5.7AI score0.00332EPSS
SaveExploits0References8Affected Software1
osv_ghsa
osv_ghsa
•added 2026/10/05 11:42 p.m.•27 views

GHSA-2823-QMQ8-RWVJ vLLM: Loose `cache_salt` validation lets a single request kill EngineCore on LMCache-MP deployments — uncaught downstream `ValueError` denial of service

Affected - Ecosystem / package: pip / vllm - Affected versions: vLLM ≤ 0.25.1 confirmed on 0.25.1, commit 752a3a504485. The lower bound predates 0.25.1; maintainers can confirm how far back the loose cachesalt validator and unguarded scheduling-path lookup reach. Summary vLLM's OpenAI-compatible...

6.5CVSS5.7AI score0.00332EPSS
SaveExploits0References5
EUVD
EUVD
•added 2026/10/05 11:42 p.m.•23 views

EUVD-2026-92846

vLLM: Loose cachesalt validation lets a single request kill EngineCore on LMCache-MP deployments — uncaught downstream ValueError denial of service...

6.5CVSS5.9AI score0.00332EPSS
SaveExploits0References4
Github Security Blog
Github Security Blog
•added 2026/10/05 11:42 p.m.•73 views

vLLM: Loose `cache_salt` validation lets a single request kill EngineCore on LMCache-MP deployments — uncaught downstream `ValueError` denial of service

Affected - Ecosystem / package: pip / vllm - Affected versions: vLLM ≤ 0.25.1 confirmed on 0.25.1, commit 752a3a504485. The lower bound predates 0.25.1; maintainers can confirm how far back the loose cachesalt validator and unguarded scheduling-path lookup reach. Summary vLLM's OpenAI-compatible...

6.5CVSS5.7AI score0.00332EPSS
SaveExploits0References5Affected Software1
NVD
NVD
•added 2026/10/05 11:17 p.m.•53 views

CVE-2026-105756

vLLM is an inference and serving engine for large language models. Prior to 0.30.0, OpenAI-compatible request models accept a non-empty cachesalt value without enforcing the character and length restrictions required by the IPCCacheServerKey consumer in LMCache-MP. On deployments using the...

6.5CVSS0.00332EPSS
SaveExploits0References4
Cvelist
Cvelist
•added 2026/10/05 10:49 p.m.•56 views

CVE-2026-105756 vLLM: Loose `cache_salt` validation lets a single request kill EngineCore on LMCache-MP deployments — uncaught downstream `ValueError` denial of service

vLLM is an inference and serving engine for large language models. Prior to 0.30.0, OpenAI-compatible request models accept a non-empty cachesalt value without enforcing the character and length restrictions required by the IPCCacheServerKey consumer in LMCache-MP. On deployments using the...

6.5CVSS0.00332EPSS
SaveExploits0References4
Vulnrichment
Vulnrichment
•added 2026/10/05 10:49 p.m.•15 views

CVE-2026-105756 vLLM: Loose `cache_salt` validation lets a single request kill EngineCore on LMCache-MP deployments — uncaught downstream `ValueError` denial of service

vLLM is an inference and serving engine for large language models. Prior to 0.30.0, OpenAI-compatible request models accept a non-empty cachesalt value without enforcing the character and length restrictions required by the IPCCacheServerKey consumer in LMCache-MP. On deployments using the...

6.5CVSS6AI score0.00332EPSS
SaveExploits0References4
CVE
CVE
•added 2026/10/05 10:49 p.m.•62 views

CVE-2026-105756

Prior to version 0.30.0 , the vLLM inference and serving engine is vulnerable to a denial of service via an uncaught ValueError . The root cause is improper input validation of the cache_salt value in OpenAI-compatible request models. On deployments using the LMCache-MP connector, providing a sal...

6.5CVSS6AI score0.00332EPSS
SaveExploits0References4Affected Software1
osv_cve
osv_cve
•added 2026/10/05 10:49 p.m.•43 views

CVE-2026-105756 vLLM: Loose `cache_salt` validation lets a single request kill EngineCore on LMCache-MP deployments — uncaught downstream `ValueError` denial of service

vLLM is an inference and serving engine for large language models. Prior to 0.30.0, OpenAI-compatible request models accept a non-empty cachesalt value without enforcing the character and length restrictions required by the IPCCacheServerKey consumer in LMCache-MP. On deployments using the...

6.5CVSS6AI score
SaveExploits0References6
Mend
Mend
•added 2026/10/05 10:49 p.m.•38 views

CVE-2026-105756

vLLM is an inference and serving engine for large language models. Prior to 0.30.0, OpenAI-compatible request models accept a non-empty cachesalt value without enforcing the character and length restrictions required by the IPCCacheServerKey consumer in LMCache-MP. On deployments using the...

7.1CVSS6AI score0.00332EPSS
SaveExploits0References7
Positive Technologies
Positive Technologies
•added 2026/10/05 12:00 a.m.•15 views

PT-2026-106266

Name of the Vulnerable Software and Affected Versions vLLM versions prior to 0.30.0 Description vLLM is an inference and serving engine for large language models. OpenAI-compatible request models accept a cache salt value without enforcing the character and length restrictions required by the...

6.5CVSS5.9AI score0.00332EPSS
SaveExploits0References16
Rapid7 Vulnerability Database (full)
Rapid7 Vulnerability Database (full)
•added 2026/10/05 12:00 a.m.•10 views

CVE-2026-105756: Improper Input Validation

vLLM is an inference and serving engine for large language models. Prior to 0.30.0, OpenAI-compatible request models accept a non-empty cachesalt value without enforcing the character and length restrictions required by the IPCCacheServerKey consumer in LMCache-MP. On deployments using the...

6.5CVSS6AI score0.00332EPSS
SaveExploits0References6
Rows per page
Query Builder