Which versions of LightLLM are affected by EUVD-2026-89990?

LightLLM, a large language model serving framework, contains a critical unauthenticated remote code execution vulnerability (CVE-2026-103395) when deployed in visual_only (multimodal) mode.

How to fix or mitigate EUVD-2026-89990?

No vendor-released patch identified at time of analysis. Within 24 hours, identify all LightLLM deployments (versions through 1.2.0) and determine which are running in visual_only mode; immediately restrict access to the RPyC port to only trusted IP addresses via firewall rules or network policies, or disable visual_only mode if not essential. Within 7 days, if visual_only mode must remain enabled, implement network segmentation to isolate the RPyC service, enforce authentication via a reverse proxy or VPN, and enable logging and monitoring for suspicious activity; also contact the vendor for a patch timeline. Within 30 days, evaluate alternative solutions or apply a custom code fix to add authentication and safe deserialization to the RPyC service, and conduct a security review of all LightLLM deployments to ensure they are not exposed to untrusted networks.

LightLLM EUVD-2026-89990

Q: What is the CVSS score of EUVD-2026-89990?

EUVD-2026-89990 has a CVSS 4.0 base score of 9.3 (Critical). CVSS vector: CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X. EPSS exploitation probability: 0.9%.

| CVE-2026-103395 CRITICAL

Deserialization of Untrusted Data (CWE-502)

2026-09-30 VulnCheck

GHSA-mw2v-h3wh-mp72

RCE Python Deserialization Lightllm

9.3

CVSS 4.0 · Vendor: VulnCheck

Severity by source

Vendor (VulnCheck) PRIMARY

9.3 CRITICAL

CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X

vuln.today AI

9.8 CRITICAL

Unauthenticated network-reachable RPyC service (PR:N/AV:N/AC:L) deserializes attacker objects for full RCE as the service account (C/I/A:H); scope unchanged as code runs within the service's own authority.

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

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

Primary rating from Vendor (VulnCheck).

CVSS VectorVendor: VulnCheck

Attack Vector

Network

Attack Complexity

Low

Privileges Required

None

User Interaction

None

Scope

Lifecycle Timeline

Metadata Corrected

Sep 30, 2026 - 15:40 vuln.today

tag: Python added

Analysis Generated

Sep 30, 2026 - 15:36 vuln.today

CVE Published

Sep 30, 2026 - 14:48 cve.org

CRITICAL 9.3

DescriptionCVE.org

LightLLM through 1.2.0 visual_only deployments expose an unauthenticated RPyC service with allow_pickle enabled that deserializes attacker-supplied arguments in the remote_infer_images method. Attackers can reach the visual RPyC port and pass objects with __reduce__ methods to execute arbitrary code with service account privileges.

Articles & Coverage 1

News Critical RCE in LightLLM Visual Mode via Unsafe Deserialization - CVE-2026-103395 Oct 01, 2026

AnalysisAI

Unauthenticated remote code execution affects LightLLM through 1.2.0 when it is deployed in visual_only (multimodal/vision) mode, which exposes an RPyC inference service that performs no authentication and unpickles the arguments passed to its remote_infer_images method. Any attacker who can reach that visual RPyC port can send a crafted object whose __reduce__ method runs on the server, achieving code execution with the privileges of the LightLLM service account and resulting in high confidentiality, integrity, and availability impact (CVSS 4.0 base 9.3; independently assessed as CVSS 3.1 AV:N/AC:L/PR:N/UI:N with high C/I/A). …

Unlock full vulnerability intelligence

Risk assessment & exploitation conditions
Attack chain visualization
Remediation with exact patch versions
Threat intelligence from 22 sources
Personal watchlist & email alerts

Continue with Google Continue with GitHub

No credit card · 7-day full trial

Attack ChainAIDerived

Hypothetical attack flow derived from CVE metadata

Access

technique details hidden

Delivery

technique details hidden

Exploit

technique details hidden

Execution

technique details hidden

Impact

technique details hidden

Access

technique details hidden

Delivery

technique details hidden

Exploit

technique details hidden

Execution

technique details hidden

Impact

technique details hidden

Vulnerability AssessmentAI

Exploitation	Requires the target to run LightLLM in 'visual_only' (multimodal vision) deployment mode, which stands up the visual RPyC service, AND for that RPyC port to be network-reachable by the attacker. … Additional conditions and limiting factors are described in the full assessment.
Risk Assessment	This is a genuine high-priority issue, not a CVSS-inflated one. … Full risk analysis with EPSS, KEV, and SSVC signal comparison available after sign-in.
Exploit Scenario	Full exploit scenario with step-by-step reproduction available after sign-in.
Remediation	No vendor-released patch was identified at time of analysis, so track the upstream repository (https://github.com/ModelTC/LightLLM) and issue https://github.com/ModelTC/LightLLM/issues/1610 for a fixed release and upgrade to that version as soon as it is tagged, using the VulnCheck advisory (https://www.vulncheck.com/advisories/lightllm-through-1.2.0-unauthenticated-remote-code-execution-via-visual-only-rpyc-service) as the reference for coverage. … Detailed patch versions, workarounds, and compensating controls in full report.