CVE-2026-103395
Deferred Deferred - Pending Action

Unauthenticated Remote Code Execution in LightLLM Visual Deployments

Vulnerability report for CVE-2026-103395, including description, CVSS score, EPSS score, affected products, exploitability, helpful resources, and attack-flow context.

Publication date: 2026-09-30

Last updated on: 2026-09-30

Assigner: VulnCheck

Description

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.

CVSS Scores

EPSS Scores

Probability:
Percentile:

Meta Information

Published
2026-09-30
Last Modified
2026-09-30
Generated
2026-09-30
AI Q&A
2026-09-30
EPSS Evaluated
N/A
NVD
EUVD

Affected Vendors & Products

Showing 2 associated CPEs
Vendor Product Version / Range
modeltc lightllm to 1.2.0 (inc)
lightllm lightllm to 1.2.0 (inc)

Helpful Resources

Exploitability

CWE
CWE Icon
KEV
KEV Icon
CWE ID Description
CWE-502 The product deserializes untrusted data without sufficiently ensuring that the resulting data will be valid.

Attack-Flow Graph

AI Quick Actions

Instant insights powered by AI
Executive Summary

CVE-2026-103395 is an unauthenticated remote code execution vulnerability in LightLLM versions up to 1.2.0. It affects visual-only deployments that expose an RPyC service with allow_pickle enabled. Attackers can connect to the visual RPyC port and pass objects with __reduce__ methods through the remote_infer_images method to execute arbitrary code with service account privileges.

Detection Guidance

Check if LightLLM visual_only deployments are running an RPyC service on port 18861 or similar. Use netstat -tulnp | grep 18861 or ss -tulnp | grep 18861 to detect listening ports. Verify if the service binds to 0.0.0.0 instead of localhost. Inspect LightLLM logs for RPyC service startup messages indicating allow_pickle enabled.

Impact Analysis

This vulnerability allows attackers to execute arbitrary code on the server with the privileges of the LightLLM service account. This could lead to full system compromise, data theft, or disruption of services. The impact is severe due to the unauthenticated nature and remote accessibility of the vulnerable service.

Mitigation Strategies

Disable the visual_only RPyC service if not required. Bind the service to localhost only instead of 0.0.0.0. Disable pickle deserialization in RPyC configuration by setting allow_pickle to False. Update LightLLM to a patched version if available. Restrict network access to the RPyC port using firewalls.

Chat Assistant

Ask questions about this CVE
Hi! I’m here to help you understand CVE-2026-103395. Ask me anything about the vulnerability, its impact, or mitigation strategies.
0/70

EPSS Chart