# @cyberjasbytes on Instagram

- **Type:** Video
- **Original URL:** https://www.instagram.com/p/DSHOr6XkXmD
- **Gondola URL:** https://gondola.cc/posts/63308048-cyberjasbytes-instagram
- **Thumbnail:** https://img.gondola.cc/tr:w-,h-,fo-auto/postThumbnails/47d49e9ddd.jpg
- **Posted:** 2025-12-11T06:42:56.000+00:00
- **Account Owner:** Jas Wong | AI & Cyber Security (@cyberjasbytes) — https://gondola.cc/cyberjasbytes

## Caption

Imagine AI telling you your bank account got hacked and to text your account number while you were on your bank’s website. Would you do it? Or know someone who would?

I easily replicated an AI attack called HashJack, a form of indirect prompt injection against browser assistants published by Cato Networks. The concerning thing is malicious instructions can be tacked onto any legitimate URL and our defenses are naturally lowered because we’re on a legit site.

Though it’s worth noting that some AI browsers fixed this issue after it was disclosed, but not all. 

👉 comment “cyber” to get a link to the HashJack article to read all the possible scenarios

——
follow @cyberjasbytes to stay in the loop on AI Security
——

#cybersecurity #promptengineering #ai #aisecurity #aihacks #hashjack #aiassistant #aibrowser 

[prompt injection, indirect prompt injection, ai awareness, cybersecurity awareness, browser security, llm security, are ai browsers safe, cybersecurity tips, cyber awareness, cyber tips, ai safety, tech awareness, ai tips, tech tips, ai research, security research, bug bounty, ai vulnerabilities, security vulnerabilities, misinformation]

## Stats

- **Views:** 184
- **Likes:** 22
- **Shares:** 0
- **Comments:** 1

## Tags

aihacks, hashjack, aibrowser, aisecurity, aiassistant, cybersecurity, promptengineering, ai

---
Copyright (c) Gondola