# New hack exploits AI hallucinations to trick agents into running malicious code — 'HalluSquatting' attack exploits a fundamental weakness in every available model

> Ever since the advent of agentic AI, security researchers have been yelling from the top of their lungs about how it's a bad idea to grant user-level permissions to an LLM — for all purposes, a program with…

- **Source:** [Tom's Hardware](https://www.tomshardware.com/tech-industry/cyber-security/hallusquatting-is-the-latest-agentic-ai-exploit-where-models-dream-up-potentially-malicious-urls-in-tool-calls-attack-exploits-a-fundamental-weakness-in-every-available-model?utm_source=gearopen.com&utm_medium=referral&utm_campaign=feed&utm_content=6a4f91a0256b604227c7e5e0)
- **Published:** 2026-07-09
- **Category:** AI & Bots
- **Tags:** #claude #ai
- **Canonical:** http://localhost:4321/a/6a4f91a0256b604227c7e5e0

---

_GearOpen's distilled summary. The full original article is published at Tom's Hardware (source link above); GearOpen does not republish the source's full text._