[NEW] [medium] vllm: Multiple vulnerabilities
An attacker can exploit multiple vulnerabilities in vllm to conduct a denial of service attack, and to disclose information.
CSIRTS triage
- What
- Multiple vulnerabilities allow denial of service attacks and information disclosure.
- Who is affected
- vllm deployments running unpatched versions.
- Urgency
- Medium severity; not currently exploited but should be patched to prevent availability and confidentiality impacts.
- Action
- Update vllm to the patched version addressing CVE-2026-71486 and CVE-2026-73560.
AI-assisted analysis generated from the source advisory — verify against the original.
⚡ Watch vllm
Get an email when a new vllm advisory drops — max one per day, one-click unsubscribe.
Details
Original advisory: https://wid.cert-bund.de/portal/wid/securityadvisory?name=WID-SEC-2026-2879
Exploitation outlook
EPSS (FIRST.org) estimates each CVE’s probability of exploitation in the next 30 days — here is the CSIRTS.com read on those numbers.
- Low exploitation riskCVE-2026-714860.34% 30-day exploitation probability — currently an unlikely target, but scores change as exploit code circulates. Riskier than 27% of all EPSS-scored CVEs.
- Low exploitation riskCVE-2026-735600.32% 30-day exploitation probability — currently an unlikely target, but scores change as exploit code circulates. Riskier than 25% of all EPSS-scored CVEs.
Referenced CVEs
| CVE | CSIRTS overview | External |
|---|---|---|
| CVE-2026-71486 | coverage & exploitation status | NVD · CVE.org |
| CVE-2026-73560 | coverage & exploitation status | NVD · CVE.org |
Same CVEs, other sources
How other CERTs, PSIRTs and databases cover the vulnerabilities in this advisory.
- mediumGHSA-4hhp-h66f-j5j7: vLLM: SSRF + arbitrary local file read in MiMoV2OmniMultiModalProcessor `_fetch_image` an…ghsa
- mediumGHSA-8737-qx52-hjff: vLLM: Derender endpoints decode caller-supplied GenerateResponse token IDs without output…ghsa
- mediumCVE-2026-73560: vLLM is an inference and serving engine for large language models. Prior to 0.26.0, the MiMoV2…nvd
- mediumCVE-2026-71486: vLLM is an inference and serving engine for large language models. Prior to 0.26.0, the /v1/co…nvd
Recent advisories for vllm
A cluster of recent advisories against the same product widens the attack surface — attackers routinely chain freshly published CVEs on one product, so review these together.
- lowCVE-2026-90713: A security flaw has been discovered in vllm-project vLLM up to 0.29.0. The affected element is…nvd · 2026-09-14
- medium[NEU] [mittel] vllm: Mehrere Schwachstellen ermöglichencert-bund · 2026-09-14
- medium[UPDATE] [mittel] vllm: Mehrere Schwachstellen ermöglichen Denial of Servicecert-bund · 2026-09-14
- mediumCVE-2026-90555: vLLM versions before 0.28.0 fail to validate audio sample rate headers in the transcription en…nvd · 2026-09-12
- mediumCVE-2026-90554: vLLM versions >=0.10.2 and <0.28.0 do not apply any audio decode-size or duration limit when e…nvd · 2026-09-12
- highCVE-2026-90553: vLLM before 0.28.0 contains a remote code execution vulnerability in the LlavaOnevision2 proce…nvd · 2026-09-12
More from CERT-Bund (BSI) Security Advisories
- medium[NEU] [mittel] Microsoft Edge: Schwachstelle ermöglicht Cross-Site Scripting2026-09-14
- medium[NEU] [mittel] Citrix Systems Workspace App Windows: Mehrere Schwachstellen ermöglichen nicht spezifizierten A…2026-09-14
- medium[NEU] [mittel] wpa_supplicant: Schwachstelle ermöglicht Umgehen von Sicherheitsvorkehrungen2026-09-14
- medium[NEU] [mittel] WP Royal Royal Elementor Addons: Schwachstelle ermöglicht Offenlegung von Informationen2026-09-14
- low[UPDATE] [niedrig] 7-Zip: Schwachstelle ermöglicht Umgehen von Sicherheitsvorkehrungen2026-09-14