Darkreading BragJack Attack Compromises Major Browser AI Assistants
Article Content
- •BragJack exploits AI assistants in five major browsers.
- •The attack allows unauthorized access to sensitive user data.
- •Two CVEs were issued, and over $20,000 in bug bounties were awarded.
The BragJack attack exploits vulnerabilities in the AI assistants of five major browsers: Google Chrome, Microsoft Edge, Opera Neon, Perplexity Comet, and Claude in Chrome. Researchers from Forever Security discovered that these vulnerabilities allow attackers to access sensitive data, control local files, and exfiltrate information through malicious browser extensions. The attack does not rely on traditional methods like prompt injection but instead leverages a critical design flaw in how browsers interact with extensions. This flaw allows untrusted extensions to hijack the communication channel with the browser's AI, leading to unauthorized actions. The research resulted in over $20,000 in bug bounties and the issuance of two CVEs: CVE-2026-0628 and CVE-2026-55945. All affected browsers have since patched the vulnerabilities. The attack poses a risk to millions of users globally who utilize these browsers with extensions installed.
Ask AI about this cluster
Answers cite the sources they use
Timeline
More articles in this cluster (19)
Following this threat?
Track Google and CVE-2026-0628 in your own feed — you're alerted when they show up in new reporting, leak sites or exploitation.
Free account · no card needed
Continue Reading
Critical RCE Vulnerability in Prefab UI Disclosed A critical remote code execution vulnerability, CVE-2026-120338, was disclosed on November 1, 2026, affecting Prefab UI's Generative UI sandbox. The flaw allows attackers to execute arbitrary Python code by exploiting Deno's permissions, enabling file read/write access on the host server. This vulnerability affects…
Chinese Actor Compromises US Government Networks via PHP2Shell Exploit A suspected Chinese nation-state actor exploited vulnerabilities in Palo Alto Networks firewalls using a chained PHP2Shell exploit and a Linux kernel privilege escalation flaw. The attack, which began on October 23, targeted 25 servers across the US, Japan, Taiwan, and South Korea, with a focus on US…