THREATOPS
THREAT OPS › Threat News › [GHSA] GHSA-g28h-2cmm-rj9x (high) — langchain-nvidia-ai-endpoints has local file disclosure through VLM image inputs

[GHSA] GHSA-g28h-2cmm-rj9x (high) — langchain-nvidia-ai-endpoints has local file disclosure through VLM image inputs

medgithub_advisoriesPublished 2026-09-24

GHSA-g28h-2cmm-rj9x Severity: high CVE: None

langchain-nvidia-ai-endpoints has local file disclosure through VLM image inputs

## Summary

`langchain-nvidia-ai-endpoints` versions before 1.4.2 accepted local filesystem paths as image inputs for Vision Language Model (VLM) requests. If an application passed attacker-controlled image input to `ChatNVIDIA` or VLM reranking APIs, an attacker could ca

Original source: https://github.com/advisories/GHSA-g28h-2cmm-rj9x