Vulnerability report

Exploited in the wild High confidence Not in CISA KEV

CVE-2025-10164

lmsys sglang update_weights_from_tensor main deserialization

Lmsys / sglang · 0.4.6

Watch CVE-2025-10164 Browse KEV feed
Severity
CVSS 6.9 · Medium
Confidence
High
Exploit status
Exploited in the wild
EPSS
0.4%
First observed
—
Last observed
—

Decision summary

What security teams need to know first

Direct answers before the deeper technical record.

What it is

CVE-2025-10164 is an unauthenticated lmsys sglang update_weights_from_tensor main deserialization. A security flaw has been discovered in lmsys sglang 0.4.6. Affected by this vulnerability is the function main of the...

Is it exploited?

Yes. Previdian tracks this CVE as a known exploited vulnerability. Confidence is high. Also confirmed by third-party sources.

Who is affected?

Lmsys / sglang 0.4.6.

What should we do?

Prioritize remediation. Validate affected assets and apply vendor fixes on an accelerated timeline.

Overview

A security flaw has been discovered in lmsys sglang 0.4.6

A security flaw has been discovered in lmsys sglang 0.4.6. Affected by this vulnerability is the function main of the file /update_weights_from_tensor. The manipulation of the argument serialized_named_tensors results in deserialization. The attack can be launched remotely. The exploit has been released to the public and may be exploited. The vendor was contacted early about this disclosure but did not respond in any way.

In lmsys sglang 0.4.6 ist eine Schwachstelle entdeckt worden. Dabei geht es um die Funktion main der Datei /update_weights_from_tensor. Die Bearbeitung des Arguments serialized_named_tensors verursacht deserialization. Der Angriff lässt sich über das Netzwerk starten. Der Exploit ist öffentlich verfügbar und könnte genutzt werden.

lmsys Affected
sglang
0.4.6
View vendor advisory (opens in new tab)
Published
09 Sep 2025
Exploitation Reported
22 Aug 2026
Attack vector
Remote
Complexity
Low
Privileges
None
User interaction
None

CVE References

Exploitation evidence

Why Previdian marks this CVE as exploited

Third-party attestation and first-party sensor observation are shown separately so teams can judge the evidence chain.

Exploited in the wild

The Shadowserver

Recorded 22 Aug 2026

A trusted third party reported exploitation.

Known exploited vulnerability sources

Per-source evidence links for KEV attestations are available through the Previdian Pro API.

Learn about Pro API access
Source Added
The Shadowserver First 2026-08-22 00:00 UTC

Detection

Operational artifacts and observed signals

Make the evidence actionable in scanner, SOC, and edge-control workflows.

Observed signals

Request targets
User-Agents
Callback hosts
0
0
0

Request targets and User-Agents available in Pro. Callback host details available in Enterprise.

Scanner coverage

No scanner integrations recorded yet.

Virtual patch status

No Previdian virtual patch is currently available. Future rules ship for ModSecurity, Cloudflare, and AWS WAF.

Learn about virtual patches →

No detection artifacts or sensor request patterns are available for this CVE yet.

Check back as sensor telemetry and scanner integrations are updated.

Risk and context

Severity, weaknesses, and research context

CVSS v4.0

6.9 Medium

Potential damage if exploited. Separate from whether attackers are using it.

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P

EPSS

0.4%

Estimated chance of exploitation in the next 30 days. Previdian's warning comes from evidence, not this score.

Weaknesses

Improper Input Validation CWE-20 Deserialization of Untrusted Data CWE-502

All CVSS Scores

CVSS v4.0 6.9 Medium

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P

CVSS v3.1 7.3 High

CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:L

CVSS v3.0 7.3 High

CVSS:3.0/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:L/E:P/RL:X/RC:R

CVSS v2.0 7.5 High

AV:N/AC:L/Au:N/C:P/I:P/A:P

Timeline

From disclosure to observed exploitation

  1. 00:00 UTC

    Added to Previdian KEV Feed

    High-confidence, third-party attested exploitation

  2. 18:32 UTC

    CVE published

    Vulnerability disclosed publicly

  3. 13:23 UTC

    CVE ID reserved

    Identifier reserved by the CNA

Pro API

Automate this intelligence

Confidence, exploit status, sensor telemetry, PoCs, scanner integrations, mentions, and tags are available programmatically for VM, SOC, and CTI workflows.

Start 14-day Pro trial Read API docs
  • Evidence confidence and provenance
  • First-party sensor telemetry
  • PoC and scanner references
  • Affected versions and enrichment
  • Automation-ready JSON delivery

GET /api/v2/pro/kevs/CVE-2025-10164

Free JSON includes basic KEV fields 
{
  "cve_id": "CVE-2025-10164",
  "confidence": "High",
  "cvss_score": 6.9,
  "cvss_estimated": false,
  "epss_score": 0.00404,
  "exploit_status": {
    "exploited_in_the_wild": true,
    "active_exploitation_observed": false
  },
  "sensor_telemetry": { "attempts": 0, "sensors": 0 }
}