# Anthropic's Unreleased AI Model Chains Flaws Into Working Chrome, Safari, Windows and Linux Exploits

DevFeed: [Anthropic's Unreleased AI Model Chains Flaws Into Working Chrome, Safari, Windows and Linux Exploits](<https://devfeed.tech/articles/anthropic-s-unreleased-ai-model-chains-flaws-into-working-chrome-safari-windows-and-linux-exploits-104365.md>)

Original publisher: [Read original article](<https://thehackernews.com/2026/10/anthropics-unreleased-ai-model-chains.html>)

Author: The Hacker News

Published: 2026-10-28T08:15:07Z

Content type: news

Language: en

Sources: [Cyber Newsroom](<https://devfeed.tech/sources/cyber-newsroom.md>)

Topics: [Security research](<https://devfeed.tech/topics/security-research.md>), [Claude](<https://devfeed.tech/topics/claude.md>), [Google Chrome](<https://devfeed.tech/topics/google-chrome.md>), [Linux](<https://devfeed.tech/topics/linux.md>), [Linux Kernel](<https://devfeed.tech/topics/linux-kernel.md>), [Windows](<https://devfeed.tech/topics/windows.md>), [Memory Safety](<https://devfeed.tech/topics/memory-safety.md>), [Android](<https://devfeed.tech/topics/android.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-model](<https://devfeed.tech/tags/ai-model.md>), [android](<https://devfeed.tech/tags/android.md>), [anthropic](<https://devfeed.tech/tags/anthropic.md>), [anthropic-ai](<https://devfeed.tech/tags/anthropic-ai.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [browser-security](<https://devfeed.tech/tags/browser-security.md>), [chrome](<https://devfeed.tech/tags/chrome.md>), [company](<https://devfeed.tech/tags/company.md>), [cve](<https://devfeed.tech/tags/cve.md>), [exploits](<https://devfeed.tech/tags/exploits.md>), [found](<https://devfeed.tech/tags/found.md>), [latest](<https://devfeed.tech/tags/latest.md>), [linux](<https://devfeed.tech/tags/linux.md>), [minutes](<https://devfeed.tech/tags/minutes.md>), [mobile-security](<https://devfeed.tech/tags/mobile-security.md>), [model](<https://devfeed.tech/tags/model.md>), [operating-system-security](<https://devfeed.tech/tags/operating-system-security.md>), [proof-of-concept](<https://devfeed.tech/tags/proof-of-concept.md>), [report](<https://devfeed.tech/tags/report.md>), [safari](<https://devfeed.tech/tags/safari.md>), [security](<https://devfeed.tech/tags/security.md>), [this](<https://devfeed.tech/tags/this.md>), [vulnerability](<https://devfeed.tech/tags/vulnerability.md>), [web-security](<https://devfeed.tech/tags/web-security.md>), [windows](<https://devfeed.tech/tags/windows.md>)

## AI overview

Anthropic says an unreleased internal model found flaws across major browsers and operating systems and chained them into five working lab proof-of-concept exploit chains; it reports no real-world attacks and says all flaws were reported and patched. The chains differ in impact: Chrome's four-bug chain reached code execution outside the browser sandbox; Safari's changed a memory value to disable same-origin protections and expose browsing data and signed-in actions; Windows' narrow denial-of-service chain crashed machines but did not steal data or run code; and Linux's chain could obtain root remotely, but required a multipath TCP listener. Anthropic reports little human involvement, but does not detail the model's work or identify it beyond an internal research model. Public records support some vendor fixes and credits, while the excerpt notes that the Windows and Linux CVE descriptions do not independently match Anthropic's account. The company says it used coordinated disclosure and points vetted security professionals to its Cyber Verification Program.

## Source excerpt

Anthropic's Unreleased AI Model Chains Flaws Into Working Chrome, Safari, Windows and Linux Exploits An unreleased Anthropic AI model found security flaws in Chrome, Safari, Android, Windows, and Linux and chained them into working proof-of-concept exploits, the company said in its latest report. The ... Read more Published Date: Oct 28, 2026 (4 hours, 51 minutes ago) Vulnerabilities has been mentioned in this article. CVE-2026-98432 CVE-2026-33096