Deserialization in Lmsys Sglang
CVE-2026-14890
SGLang uses an expert-parallel backup subsystem that exposes a ZeroMQ PULL socket on a routable network interface that does not contain authentication or deserialization safeguards, allowing an attacker to provide a malicious pickle file that results in unauthenticated remote code execution when the feature is enabled and the service is reachable over the network.
Vulnerability class: Insecure Deserialization
Published · last modified .
CVSS v3 metric
CVSS v3 base score 9.1 (Critical). Vector: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:N.
EPSS exploit prediction
EPSS: 0.010 (62.4th percentile), scored .
Very low probability of exploitation in the next 30 days; routine patching cadence is appropriate. 62nd percentile — 62.4% of CVEs in the catalogue have a lower EPSS than this one. How to read EPSS.
EPSS over last 30 days · oldest: 0.010 · newest: 0.010 · change: 0.000
Affected products
- Lmsys Sglang
- Sglang — versions 0
Weakness classification (CWE)
References
- github.com/sgl-project/sglang/blob/main/python/sglang/srt/elastic_ep/expert_bac… (Product)
- vince.cert.org/vuls/id/326070 (Permissions Required)
- af854a3a-2127-422b-91ae-364da2661108 (Third Party Advisory, Mitigation)
Frequently asked questions
- What is CVE-2026-14890?
- CVE-2026-14890 is a critical-severity vulnerability in Lmsys Sglang, classified under Deserialization of Untrusted Data. CVSS score: 9.1/10. Published 2026-07-16.
- How severe is CVE-2026-14890?
- Critical severity. CVSS v3 base score is 9.1 out of 10.