#5
The Hacker News
general
August 25, 2026 at 14:07 UTC
A Malicious Webpage Could Poison Your Local AI Model Behind NVIDIA NemoClaw
By [email protected] (The Hacker News)
AI Summary
Oasis Security disclosed a vulnerability in NVIDIA NemoClaw (OpenClaw) that allows an attacker-controlled webpage to gain unauthenticated access to a local Ollama API instance, enabling persistent poisoning of AI agent model instructions. The attack requires no authentication and can persist hidden instructions inside the model itself, representing a novel supply-chain-style risk for AI agent deployments.
Relevance score: 80.0/100
Sponsored
Protect Your Business
Expert cybersecurity solutions to safeguard your organization from evolving threats.
Get Protected →