vLLM is an inference and serving engine for large language models (LLMs). Starting in version 0.10.1 and prior to version 0.18.0, two model implementation files hardcode trust_remote_code=True when loading sub-components, bypassing the user's explicit --trust-remote-code=False security opt-out. This enables remote code execution via malicious model repositories even when the user has explicitly disabled remote code trust. Version 0.18.0 patches the issue.
CVE-2026-27893
HIGHPre-NVD 8.88.8
EchelonGraph scoreLOW confidence
This high-severity CVE scores 8.8 under a secondary CVSS source (NVD's own analysis pending). EPSS exploit-prediction score not yet available (the EPSS model rescores nightly; freshly-published CVEs typically appear within 48 hours). GitHub Security Advisory data not yet ingested — confidence will rise once GHSA publishes (typical lag: hours to days for open-source ecosystem CVEs; never for infrastructure-only CVEs).
Triggered by: NVD CVSS baseline
Sources: secondary
8.8
EchelonGraph verdictPlan a fixSerious severity, but no confirmed exploitation yet.
- High severity, but no confirmed exploitation yet
CISA-KEV: Not listedEPSS: 1%CVSS: 8.8Exploit: NoneExposed: 0
A fix is available — apply it.
- CVSS v3
- 8.8
- EG Score
- 8.8(low)
- EPSS
- 68.5%
- KEV
- Not listed
Published
March 27, 2026
Last Modified
June 30, 2026
Advisory Details (10)
Auto-updated Jun 30, 2026Patch available. Sources: github_commit, github_pr, github, redhat.
redhat Patch Available
redhat Patch Available
redhat Patch Available
redhat Patch Available
redhat Patch Available
redhat Patch Available
redhat Patch Available
github Patch Available
Hardcoded trust_remote_code=True in NemotronVL and KimiK25 bypasses user security opt-out · Advisory · vllm-project/vllm · GitHub
https://github.com/vllm-project/vllm/security/advisories/GHSA-7972-pg2x-xr59github_pr
[Security] Respect user trust_remote_code setting in NemotronVL and KimiK25
Fix merged in vllm-project/vllm PR #36192 on 2026-03-06 — awaiting tagged release
https://github.com/vllm-project/vllm/pull/36192github_commit
commit 00bd08edeee5 (vllm-project/vllm)
Fix landed in vllm-project/vllm commit 00bd08edeee5 — awaiting tagged release
https://github.com/vllm-project/vllm/commit/00bd08edeee5dd4d4c13277c0114a464011acf72Vendor Advisories for CVE-2026-27893(9)
These vendors published their own advisory mentioning this CVE — often with vendor-specific remediation steps + affected product lists not in NVD.
- RHSA-2026:24977Red Hat Product SecurityHigh
Red Hat Security Advisory: RHOAI 2.25.7 - Red Hat OpenShift AI
- RHSA-2026:19725Red Hat Product SecurityHigh
Red Hat Security Advisory: Red Hat AI Inference Server 3.2.2 (ROCm)
- RHSA-2026:19724Red Hat Product SecurityHigh
Red Hat Security Advisory: Red Hat AI Inference Server 3.2.2 (CUDA)
- RHSA-2026:19712Red Hat Product SecurityCritical
Red Hat Security Advisory: RHOAI 3.3.3 - Red Hat OpenShift AI
- RHSA-2026:10141Red Hat Product SecurityHigh
Red Hat Security Advisory: Red Hat Enterprise Linux AI 3.3.1
- RHSA-2026:10140Red Hat Product SecurityHigh
Red Hat Security Advisory: Red Hat Enterprise Linux AI 3.3.1
- RHSA-2026:8748Red Hat Product SecurityHigh
Red Hat Security Advisory: Red Hat AI Inference Server Model Optimization Tools 3.3.1 (CUDA)
- RHSA-2026:8746Red Hat Product SecurityHigh
Red Hat Security Advisory: Red Hat AI Inference Server 3.3.1 (CUDA)
- +1 more
Patch Availability(9)
| Vendor / Ecosystem | Fixed in / Patch | Released | Source |
|---|---|---|---|
| redhat | rhoai/odh-vllm-gaudi-rhel9:1780069069 | 2026-06-10 | redhat |
| redhat | rhoai/odh-vllm-gaudi-rhel9:1778600187 | 2026-05-20 | redhat |
| redhat | rhaiis/vllm-rocm-rhel9:1779223651 | 2026-05-20 | redhat |
| redhat | rhaiis/vllm-cuda-rhel9:1779223654 | 2026-05-20 | redhat |
| redhat | rhelai3/disk-image-cuda-rhel9:1776938871 | 2026-04-23 | redhat |
| redhat | rhelai3/bootc-rocm-rhel9:1776773505 | 2026-04-23 | redhat |
| redhat | rhaiis/vllm-rocm-rhel9:1775680262 | 2026-04-17 | redhat |
| redhat | rhaiis/vllm-cuda-rhel9:1775680192 | 2026-04-17 | redhat |
| redhat | rhaiis/model-opt-cuda-rhel9:1775749857 | 2026-04-17 | redhat |
Patches are aggregated from vendor advisories (Red Hat, Microsoft, Cisco, GitHub) and package ecosystems (OSV, GHSA). Multiple rows for the same upstream release have been deduplicated.
Affected Packages
(1 across 1 ecosystem)
PyPI(1)
| Package | Vulnerable range | Fixed in | Dependents |
|---|---|---|---|
| vllm | 0.10.1 ... 0.17.1 (15 versions) | 0.18.0 | — |
Weakness Classification(2)
MITRE Common Weakness Enumeration — the root-cause categories this CVE belongs to.
Data Freshness Timeline
(refreshed 11× in last 7d / 14× in last 30d)
Each row is a source pipeline that fetched or updated this CVE on that date, with what changed. For example, "NVD update" means NVD published or revised its analysis for this CVE; "MITRE cvelistV5" means we ingested or refreshed it from the CNA feed. Most recent first.
- 2026-07-06 16:27 UTCEPSS rescore
- 2026-07-06 16:27 UTCEPSS rescore
- 2026-07-06 02:23 UTCEPSS rescore
- 2026-07-06 02:23 UTCEPSS rescore
- 2026-07-05 02:30 UTCEPSS rescore
- 2026-07-04 06:31 UTCEPSS rescore
- 2026-07-04 06:31 UTCEPSS rescore
- 2026-07-01 15:06 UTCEPSS rescore
- 2026-07-01 15:06 UTCEPSS rescore
- 2026-06-30 23:22 UTCEPSS rescore
- 2026-06-30 23:22 UTCEPSS rescore
- 2026-06-30 04:37 UTCEG score recompute
- 2026-06-30 04:37 UTCVendor advisory
- 2026-06-30 04:32 UTCNVD updatefirst tracked
Related CVEs(same vendor + same CWE)
Same vendor
10 shownredhat
Same CWE
10 shownCWE-693 · CWE-501
- CVE-2021-32835EG 9.9EPSS 90%CRITICAL
- CVE-2019-10328EG 9.9CRITICAL
- CVE-2019-1003030EG 9.9 KEVEPSS 99%CRITICAL
- CVE-2022-1799NVD 5.7EG 9.8MEDIUM
- CVE-2021-27497NVD 6.5EG 9.8MEDIUM
- CVE-2020-10887EG 9.8EPSS 90%CRITICAL
- CVE-2017-3197EG 9.8EPSS 92%CRITICAL
- CVE-2018-9318EG 9.8EPSS 90%CRITICAL
- CVE-2018-9311EG 9.8CRITICAL
- CVE-2013-2465EG 9.8 KEVEPSS 100%CRITICAL
Frequently asked(5)
What is CVE-2026-27893?
CVE-2026-27893 is a high vulnerability published on March 27, 2026. vLLM is an inference and serving engine for large language models (LLMs). Starting in version 0.10.1 and prior to version 0.18.0, two model implementation files hardcode trustremotecode=True when loading sub-components, bypassing the user's explicit --trust-remote-code=False security opt-out. This…
When was CVE-2026-27893 disclosed?
CVE-2026-27893 was first published in the National Vulnerability Database on March 27, 2026, with the most recent update on June 30, 2026. EchelonGraph re-ingests CVE updates from NVD on a 2-hour cycle, so this page reflects the latest published state.
Is CVE-2026-27893 actively exploited?
CVE-2026-27893 is not currently on CISA's Known Exploited Vulnerabilities catalog. FIRST EPSS estimates a 68.5% percentile likelihood of exploitation in the next 30 days — higher percentiles indicate greater predicted risk.
What is the CVSS score of CVE-2026-27893?
CVE-2026-27893 has a CVSS v3 base score of 8.8 (NVD). The EG score is currently aggregating — additional source signals are being incorporated as they become available..
How do I remediate CVE-2026-27893?
Patch to the fixed version published by the affected vendor. Where vendor advisories exist for CVE-2026-27893, EchelonGraph cross-links them in the Vendor Advisories panel below — those typically contain the canonical remediation steps, fixed version numbers, and any vendor-specific mitigations.
Dependency Blast Radius
See which npm, PyPI, Go, and Maven packages are affected by CVE-2026-27893
Is Your Infrastructure Affected by CVE-2026-27893?
EchelonGraph automatically scans your cloud infrastructure and maps CVE exposure using blast radius analysis.