vLLM Vulnerabilities

pipLLM Inference

AI Threat Alert tracks 188 known vulnerabilities in vLLM, 13 rated critical — an AI/ML llm inference in the pip ecosystem. Each CVE includes CVSS severity, EPSS exploit probability, patch status, and CISO-grade analysis.

61
Risk Score
188
Total CVEs
13
Critical
pip
Ecosystem
Nov 18, 2026
Last CVE
26%
Patch Rate
37d
Avg Time to Patch
96,942 stars24,834 forks10,395 issues93 dependentsLast push Nov 15, 2026
View on GitHub

Known Vulnerabilities(188 total, page 3 of 8)

SeverityCVE IDSummaryCVSSPublished
LOWCVE-2025-6170libxml2: xmllint shell stack overflow causes local crash2.5Jun 16, 2025MEDIUMCVE-2026-58015GLib: D-Bus cookie path traversal leaks local files5.9Jun 30, 2026HIGHCVE-2026-58014GLib: off-by-one in key file parser enables DoS7.3Jun 30, 2026MEDIUMCVE-2026-58011GLib: out-of-bounds read in g_date_time_get_ymd6.5Jun 30, 2026HIGHCVE-2026-16118xdgmime: heap overflow in magic file parsing7.1Jul 17, 2026MEDIUMCVE-2026-58013GLib: buffer over-read in line reader causes DoS6.5Jun 30, 2026MEDIUMCVE-2026-58012GLib: g_regex_replace RAW over-read leaks bytes, DoS6.5Jun 30, 2026MEDIUMCVE-2026-58010GLib: off-by-one read in GVariant causes DoS6.5Jun 30, 2026MEDIUMCVE-2026-15588GLib GDBus: unbounded auth line enables local DoS5.3Jul 20, 2026HIGHCVE-2026-103241vLLM: Gemma4 parser flaw allows remote DoS7.5Sep 30, 2026MEDIUMCVE-2026-100654vLLM: out-of-range stop_token_ids crash EngineCore (DoS)6.5Sep 26, 2026MEDIUMCVE-2026-100653vLLM: revision pin bypass loads unpinned HF artifacts6.5Sep 26, 2026HIGHCVE-2026-100652vLLM: invalid stop_token_ids crash EngineCore (DoS)7.5Sep 26, 2026MEDIUMCVE-2026-100651vLLM: prompt-length bypass on disagg endpoint causes DoS6.5Sep 26, 2026MEDIUMCVE-2026-100650vLLM: unbounded media fetch enables pre-inference DoS6.5Sep 26, 2026MEDIUMCVE-2026-100649vLLM: sampler subclass bypass exhausts GPU memory5.3Sep 26, 2026MEDIUMCVE-2026-100648vLLM: audio size limit bypass enables resource exhaustion5.3Sep 26, 2026HIGHCVE-2026-100647vLLM: unbounded cache_salt allows unauthenticated DoS7.5Sep 26, 2026CRITICALCVE-2026-61732Decepticon: ChatML token injection yields sandbox RCE10.0Sep 24, 2026HIGHCVE-2026-94627vLLM: Mooncake KV cache leak enables GPU DoS7.5Sep 21, 2026HIGHCVE-2026-94626vLLM: unvalidated tp_size allows memory-exhaustion DoS7.5Sep 21, 2026MEDIUMCVE-2026-94625vLLM: Mooncake placeholder leak enables slow DoS5.3Sep 21, 2026HIGHCVE-2026-94624vLLM: unbounded P2P KV sessions crash EngineCore (DoS)7.5Sep 21, 2026HIGHCVE-2026-94623vLLM: NIXL multi-prompt assertion crash causes DoS7.5Sep 21, 2026HIGHCVE-2026-94622vLLM: NIXL KeyError crashes decode engine remotely7.5Sep 21, 2026

Showing 51–75 of 188

Frequently asked questions

What is vLLM?

vLLM is an AI/ML llm inference tracked by AI Threat Alert for security vulnerabilities in the pip ecosystem.

How many known vulnerabilities does vLLM have?

vLLM has 188 known CVEs, 13 of them critical, tracked from NVD and GitHub Advisory.

Which ecosystem is vLLM distributed in?

vLLM is distributed via the pip ecosystem and categorized as llm inference.

Where does the vLLM vulnerability data come from?

Vulnerability data is sourced from NVD and GitHub Advisory, enriched with CVSS, EPSS, exploit signals, and patch status for each CVE.

How do I assess the risk of vLLM?

Review each CVE below — every entry shows CVSS severity, EPSS exploit probability, exploitation signals, and whether a patched version is available.

Monitor vLLM in your stack

Get instant alerts when new vulnerabilities affect vLLM. CISO analysis, ATLAS technique mappings, and compliance reports included.

Start Monitoring