CVE-2026-41523Disclosure(vllm / vllm)

LOWCVSS 7.5 · HIGH

Signal is active with 1 mentions in latest observed window

Immediate actions

Patch vllm vllm systems immediately

Recommended action window: Monitor and triage in normal cycle

NVD description

vLLM is an inference and serving engine for large language models (LLMs). Prior to 0.22.0, an assert-based security check in vLLM's activation function loading allows any unauthenticated attacker to achieve arbitrary code execution on the server by publishing a malicious HuggingFace model, when vLLM runs in Python optimized mode (python -O or PYTHONOPTIMIZE=1). This vulnerability is fixed in 0.22.0.

vllm CVEs Back to dashboard

0.5/ 10 priority

Sources & remediation

Patch / fix

https://github.com/vllm-project/vllm/commit/b3c7ffcab82c2439726f8cb213800f6f38c023d3

Exploit / PoC references

https://github.com/vllm-project/vllm/security/advisories/GHSA-q8gq-377p-jq3r

Vendor / third-party advisories

https://huntr.com/bounties/dcb05b04-e625-41e7-adbc-bbae0cc2d64c

Other references

Weakness type (CWE)

CWE-94CWE-617

Priority

LOW

Exploitation

NONE

PoC

YES

Patch

AVAILABLE

Momentum

STABLE

Are you affected?

If you run products in this scope, you should treat this CVE as relevant to your environment.

vllm

Threat summary

Patch or workaround signal is available
4 mentions across 3 observed days
Momentum state: stable

What's happening

Patch or workaround mentioned in 1 signal
Technical details provided in 4 signals
Disclosure: 3 classified signals
Peaked 2d ago at 2 mentions (2026-06-22); latest day: 1
4 total mentions across 3 days

Affected systems

Vendors

vllm

Products

vllm

Deep dive

Activity timeline4 mentions / 3d

Signal classification2 categories

Disclosure

375.0%

Patch

125.0%

Referenced assets3 URLs

By indicator

Classification over time

Date	Total	Labels

2026-06-22	2	Disclosure2
2026-06-23	1	Patch1
2026-06-24	1	Disclosure1

Full discourse4 posts

Cyber Threat Observatory | Alan Turing Institute@TuringCyberObs90d
Disclosure
CVE-2026-41523 — vLLM The vulnerability can allow code execution when malicious Hugging Face models are loaded while Python optimised mode is enabled. Full analysis: https://github.com/alan-turing-institute/cyber-threat-observatory/blob/main/reports/2026-06-22/TIER_2_CVE-2026-41523.md #CyberSecurity #AIsecurity #VulnerabilityManagement #CyberObs
Post summary
The report describes CVE-2026-41523 in vLLM as a code‑execution flaw triggered by malicious Hugging Face models in Python's optimized mode, but provides no PoC, exploit, patch, or evidence of active exploitation.
0000043
55 followersView on X
DFIR Lab@DFIR_Lab91d
Patch
🚨 HIGH: CVE-2026-41523 (CVSS 7.5) - vLLM inference engine vulnerable to RCE via malicious HuggingFace models when running in Python optimized mode. Patch to v0.22.0 immediately. #CVE #Vulnerability #PatchNow https://t.co/Q72wgpKSKF
Post summary
High‑severity RCE vulnerability in vLLM inference engine; immediate patch to v0.22.0 is recommended.
0000050
50 followersView on X
Infoflowcloud@infoflowcloud91d
Disclosure
🚨*CVE* CVE-2026-41523 vLLM is an inference and serving engine for large language models (LLMs). Prior to 0.22.0, an assert-based security check in vLLM's activation function loading allows… https://www.cve.org/CVERecord?id=CVE-2026-41523 ----- Traducción: CVE-2026-41523 vLL… http://infoflow.cloud`
Post summary
The tweet announces CVE-2026-41523, noting an assert-based security flaw in vLLM prior to version 0.22.0, but does not provide a PoC, exploit, or patch details.
0000033
88 followersView on X
CVE@CVEnew91d
Disclosure
CVE-2026-41523 vLLM is an inference and serving engine for large language models (LLMs). Prior to 0.22.0, an assert-based security check in vLLM's activation function loading allows… https://www.cve.org/CVERecord?id=CVE-2026-41523
Post summary
The CVE-2026-41523 entry reports a security flaw in vLLM’s activation function loading before version 0.22.0, with technical details but no PoC, patch, or evidence of exploitation.
00000744
57.7K followersView on X

CPE platform detail1 entries

Part:

Vendor:

Product:

Version:

Target SW:

Target HW:

1 of 1 entries

Part	Vendor	Product	Version	Target SW	Target HW
App	vllm	vllm	-	-	-

Explore more

Back to dashboard vllm CVEs vllm vulnerabilities

What researchers are saying

CVE-2026-41523

DFIR Lab[verified]@DFIR_LabPatch

High‑severity RCE vulnerability in vLLM inference engine; immediate patch to v0.22.0 is recommended.

Cyber Threat Observatory | Alan Turing Institute@TuringCyberObsDisclosure

The report describes CVE-2026-41523 in vLLM as a code‑execution flaw triggered by malicious Hugging Face models in Python's optimized mode, but provides no PoC, exploit, patch, or evidence of active exploitation.

Infoflowcloud@infoflowcloudDisclosure

The tweet announces CVE-2026-41523, noting an assert-based security flaw in vLLM prior to version 0.22.0, but does not provide a PoC, exploit, or patch details.

CVE@CVEnewDisclosure

The CVE-2026-41523 entry reports a security flaw in vLLM’s activation function loading before version 0.22.0, with technical details but no PoC, patch, or evidence of exploitation.