ainewsblitz.com

Breaking

LayerX Shows How to Fool 6 AI Browsers Into Stealing GitHub Keys via a 'Game'

LayerX Shows How to Fool 6 AI Browsers Into Stealing GitHub Keys via a 'Game'
  • Security
  • AI Agents

Researchers at security firm LayerX have demonstrated an attack called "BioShocking" that bypasses safety controls by tricking AI browsers/agents into believing they are "playing a game," causing them to copy and leak a user's GitHub SSH credentials. Proof-of-concept (PoC) attacks succeeded against six agentic browsers/extensions, including ChatGPT Atlas, Comet, and Claude. The details were published in late June 2026 on LayerX's technical blog.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.

$20
Read this article
$29/month
Unlimited — all 28,543 articles, the full archive, and comprehension quizzes
Save 72%
$98/year
≈ $8.17/month
Unlimited, billed once a year