Topic

vLLM: news, advisories and analysis

11 articlesUpdated 18 Oct 2026

CyberSecureToday covers vLLM across 11 articles published between 20 Sep 2026 and 18 Oct 2026. The coverage includes 10 vulnerability advisories and 1 analysis or guidance piece. 10 advisories are rated critical or high severity.

Stories on vLLM most often intersect with AI Security, PagedAttention and GPU Security, with Technology, Cloud & SaaS the most affected sector. The most recent story, “LMCache Insecure Deserialization Flaw Exposes vLLM AI Inference Nodes to Unauthenticated Root RCE”, was published 18 Oct 2026. Every article is listed below, newest first, with the affected products, severity and remediation guidance in each piece.

Latest coverage

ThreatsCloud & SaaSCritical
LMCache Insecure Deserialization Flaw Exposes vLLM AI Inference Nodes to Unauthenticated Root RCE
A critical unauthenticated deserialization flaw in LMCache (CVE-2026-105192) allows remote attackers to execute arbitrary code with root privileges on vLLM clusters.
Aisha Noor18 Oct 202610 min read
NewsCloud & SaaSCritical
vLLM Multi-LoRA Serving Engine Flaw Enables Cross-Tenant Adapter Memory Extraction and Model Poisoning (CVE-2026-78430)
vLLM patches a high-severity vulnerability (CVE-2026-78430) in multi-LoRA serving allowing concurrent tenants to extract proprietary adapter weights from GPU memory.
CST Editorial Desk13 Oct 20268 min read
ThreatsCloud & SaaSCritical
vLLM Continuous Batching Engine Vulnerability Enables Cross-Tenant KV-Cache Extraction and Host Memory Corruption (CVE-2026-72110)
vLLM issues critical advisory CVE-2026-72110 for its continuous batching inference engine, patching a PagedAttention memory calculation bug that leaked LLM KV-cache tokens.
CST Editorial Desk11 Oct 20268 min read
ThreatsCloud & SaaSHigh
vLLM Distributed KV-Cache Memory Deserialization Flaw Enables Arbitrary Model Weight Extraction
A high-severity deserialization vulnerability in vLLM's distributed PagedAttention KV-cache transfer protocol allows attackers to compromise GPU worker nodes.
CST Editorial Desk10 Oct 20268 min read
ThreatsCloud & SaaSCritical
vLLM Flaw Lets Unauthenticated Attackers Poison KV Caches and Hijack Model Outputs
A critical flaw in vLLM distributed inference engine (CVE-2026-43109) allows unauthenticated network actors to poison PagedAttention KV caches and hijack model outputs.
Aisha Noor7 Oct 20268 min read
ThreatsCloud & SaaSHigh
vLLM EngineCore Guided Decoding State Failure Triggers Multi-Tenant GPU Cluster Denial of Service (GHSA-85xf-c7hm-whqw)
A high-severity flaw in vLLM EngineCore guided decoding allows crafted JSON Schema requests to trigger non-recoverable state crashes across multi-tenant GPU workers.
Mei-Lin Chen30 Sep 202610 min read
ThreatsCloud & SaaSHigh
Malicious PyPI Package Masquerades as vLLM Process to Exfiltrate Vector Database Weights (GHSA-rghm-9c3j-wc97)
A sophisticated typosquatted Python package targets AI engineering pipelines, masquerading as vLLM inference worker daemons to exfiltrate proprietary model weights.
CST Editorial Desk30 Sep 202610 min read
ThreatsCloud & SaaSCritical
vLLM CUDA Kernel Integer Overflow Leaks Cross-Tenant Inference Prompts
An integer overflow flaw in vLLM PagedAttention CUDA kernels allows concurrent inference requests to read sensitive prompt tokens across multi-tenant GPU memory.
Mei-Lin Chen27 Sep 20269 min read
ResearchCloud & SaaS
vLLM Multimodal Special Token Injection Triggers Remote Engine Crash and Service Denial CVE-2026-44222
High-severity flaw CVE-2026-44222 in vLLM allows remote unauthenticated users to crash distributed GPU inference clusters via crafted multimodal tokens.
CST Editorial Desk23 Sep 202615 min read
ThreatsCloud & SaaSCritical
vLLM Inference Server Authentication Bypass Flaw CVE-2026-48746 Exposes Enterprise Model Endpoints Without API Keys
Critical CVSS 9.1 flaw CVE-2026-48746 in vLLM allows unauthenticated remote attackers to bypass API key verification and access model completions due to URL normalization defects.
Mei-Lin Chen21 Sep 202614 min read
ThreatsCloud & SaaSCritical
vLLM Video Parser Buffer Overflow Enables Remote Root Code Execution
vLLM maintainers and NIST disclose critical CVSS 9.8 flaw CVE-2026-22778 in multimodal video processing allowing unauthenticated remote root takeover of GPU clusters.
Daniel Okafor20 Sep 202614 min read