Is CVE-2026-103243 exploited in the wild?

No. Junglewise Threat Intelligence has no report of CVE-2026-103243 being exploited in the wild.

How severe is CVE-2026-103243?

Medium severity, with a CVSS score of 5.8, published 30 September 2026.

Which products and versions are affected?

ModelTC LightLLM through 1.2.0.

Is there a fix for CVE-2026-103243?

No fixed version is recorded yet. Check the vendor's advisory in the references below. If you run it, Junglewise drafts a mitigation that fits your setup and applies it once you approve, with senior security engineers behind it.

How is CVE-2026-103243 attacked?

It is attacked over the network, in a software library.

Am I affected by CVE-2026-103243?

Check whether you run ModelTC LightLLM at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

CVE-2026-103243: LightLLM server-side request forgery in multimodal endpoints

CVE-2026-103243 · Severity: medium · CVSS 5.8 · Published 2026-09-30

LightLLM is a Python-based framework for serving large language models, including multimodal models that process images and audio. An unauthenticated attacker can exploit the image_url and audio_url parameters in the inference API to make the server fetch arbitrary internal resources, potentially reading sensitive data from internal networks or probing the internal network topology through error messages.

Technologies: ModelTC LightLLM. Vendors: ModelTC.

Quick answers

Is CVE-2026-103243 exploited in the wild?: No. Junglewise Threat Intelligence has no report of CVE-2026-103243 being exploited in the wild.
How severe is CVE-2026-103243?: Medium severity, with a CVSS score of 5.8, published 30 September 2026.
Which products and versions are affected?: ModelTC LightLLM through 1.2.0.
Is there a fix for CVE-2026-103243?: No fixed version is recorded yet. Check the vendor's advisory in the references below. If you run it, Junglewise drafts a mitigation that fits your setup and applies it once you approve, with senior security engineers behind it.
How is CVE-2026-103243 attacked?: It is attacked over the network, in a software library.
Am I affected by CVE-2026-103243?: Check whether you run ModelTC LightLLM at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

Technical details

LightLLM fails to validate or filter image_url and audio_url parameters in multimodal endpoints, allowing unauthenticated server-side request forgery. The fetch_resource() function in multimodal_utils.py accepts arbitrary URLs without scheme allowlisting, private address filtering, or origin validation; fetched content is passed to vision models for processing, and HTTP errors are returned to the client with details that enable port/path probing. Attackers can read internal image resources via model output or enumerate internal services through distinct error messages for 404 responses, closed ports, and non-image responses.

Affected products

ModelTC LightLLM through 1.2.0

Timeline

2026-09-30: disclosed

References

This page as Markdown. Free to reuse under CC BY 4.0 with a link back.