langchain-nvidia-ai-endpoints has local file disclosure through VLM image inputs
- Product
- langchain-nvidia-ai-endpoints
- Affected
- langchain-nvidia-ai-endpoints <= 1.4.1
- Fixed in
- 1.4.2
Last verified against sources: October 26, 2026
Am I affected?
langchain-nvidia-ai-endpoints: affected langchain-nvidia-ai-endpoints <= 1.4.1. Fixed in 1.4.2.
- Check your version: run
pip show langchain-nvidia-ai-endpoints. Versions <= 1.4.1 of langchain-nvidia-ai-endpoints are affected. [1]
What to do now
- Upgrade to the fixed version: run
pip install --upgrade "langchain-nvidia-ai-endpoints>=1.4.2"(first patched version 1.4.2). [1] - Upgrade to langchain-nvidia-ai-endpoints >= 1.4.2. [1]
- If upgrading is not immediately possible, reject local filesystem paths in user-controlled VLM image inputs and only allow trusted remote URLs, data URIs, or known safe asset/file IDs. [1]
- Run the application with least-privilege filesystem access so the application process cannot read files it does not need. [1]
General practice, not from the source: if the affected software could read tokens or credentials, rotate them after upgrading; pin versions and review changes before updating tools your agents run.
What happened
- Versions of langchain-nvidia-ai-endpoints before 1.4.2 accepted local filesystem paths as image inputs for Vision Language Model (VLM) requests. [1]
- If an application passed attacker-controlled image input to ChatNVIDIA or VLM reranking APIs, a file readable by the application process could be read and included in the outbound request to the configured NVIDIA/NIM model endpoint. [1]
Impact
What was done
Sources
- GHSA-g28h-2cmm-rj9x — GitHub Advisory Database
Every statement above was checked against these sources before publishing. Spotted an error? Tell us and we will correct it with a dated note.
Share advisory brief with your team
Copy a plain-text summary for Slack, GitHub issues, or team security channels.
langchain-nvidia-ai-endpoints has local file disclosure through VLM image inputs
Product: langchain-nvidia-ai-endpoints | Affected: langchain-nvidia-ai-endpoints <= 1.4.1 | Fixed in: 1.4.2
Summary: Versions of langchain-nvidia-ai-endpoints before 1.4.2 accepted local filesystem paths as image inputs for Vision Language Model (VLM) requests.
Source & remediation on AgentArmorHQ: https://agentarmorhq.com/incidents/2026/09/ghsa-g28h-2cmm-rj9x-langchain-nvidia-ai-endpointsCorrections
Report an error or request a correction. We review requests within 5 business days and add a dated correction note.
Discussions for langchain-nvidia-ai-endpoints
Ask questions or share configuration tips. Governed by our Community Rules & Terms.
No posts match this filter yet.
Be the first to ask a question or share a working tip.