AIThe Hacker Newsabout 2 hours ago
NVIDIA NemoClaw flaw enables AI model poisoning
1 min read
Oasis Security disclosed CVE-2026-65105 in NVIDIA NemoClaw, letting a malicious webpage take unauthenticated control of the local Ollama server and inject hidden instructions into the model. The flaw bypasses authentication via DNS rebinding, enabling model poisoning that survives chat resets.
Level
Hype check
Tap to vote and see what everyone thinks.
In this storyNvidia
