<!-- mobian-agent-page publisher="dailydev" canonical="https://daily.dev/posts/claude-builds-exploits-across-major-browsers-oses-hqofelevw" -->

---
title: Claude Builds Exploits Across Major Browsers & OSes
description: An unreleased version of Claude found and chained vulnerabilities into five working exploits against Chrome, Safari, Android, Windows, and Linux. One attack...
canonical: https://daily.dev/posts/claude-builds-exploits-across-major-browsers-oses-hqofelevw
twitter:card: summary_large_image
twitter:site: @dailydotdev
og:type: website
og:site_name: daily.dev
og:title: Claude Builds Exploits Across Major Browsers & OSes | daily.dev
og:description: An unreleased version of Claude found and chained vulnerabilities into five working exploits against Chrome, Safari, Android, Windows, and Linux. One attack...
og:url: https://daily.dev/posts/claude-builds-exploits-across-major-browsers-oses-hqofelevw
og:image: https://api.daily.dev/og/posts/HqofelEvw.png
og:image:alt: Claude Builds Exploits Across Major Browsers & OSes
og:image:width: 1200
og:image:height: 630
og:locale: en
---

> ## Documentation Index
> Fetch the complete documentation index at: https://daily.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# Claude Builds Exploits Across Major Browsers & OSes

**[Dark Reading](https://daily.dev/sources/dr)** · 6 min read · 0 upvotes · 0 comments

## Summary

An unreleased version of Claude found and chained vulnerabilities into five working exploits against Chrome, Safari, Android, Windows, and Linux. One attack achieved remote code execution with root privileges on Linux servers via a zero-click kernel networking flaw. Another extracted photos from a locked Google Pixel 9 by chaining three Android USB-handling bugs. Anthropic reported 62 vulnerabilities to Google, all patched by Chrome 156, and says all affected vendors have fixed the issues. Black Duck's Andrew Bolster notes the report gives no success rates or human-hours per chain, and VulnCheck found only 1.3% of AI-discovered vulnerabilities are exploited in the wild.

## Full article

daily.dev links to this article rather than hosting it. Read it at the original source: <https://www.darkreading.com/application-security/claude-model-exploits-major-browsers-oses>

## Questions this post answers

### What vulnerabilities did Claude find and exploit across Chrome, Safari, Android, Windows, and Linux?

An unreleased Claude model generated five working exploits: a one-click Chrome RCE chaining four bugs in the JS engine, memory protections, process isolation, and federated sign-in; a Safari same-origin-policy bypass chaining four WebKit flaws; an Android exploit extracting data from a locked Pixel 9 via USB-handling bugs; a Windows kernel HTTP/3 denial-of-service; and a zero-click Linux kernel networking RCE with root privileges, all now patched.

_Developers tracking AI-driven vulnerability discovery can follow coverage like this on daily.dev._

### How many vulnerabilities did Anthropic report to Google's Chrome team using Claude?

Anthropic reported 62 vulnerabilities to Google, all of which received CVE identifiers, were rated high or critical severity, and were patched by Chrome version 156. This was part of Anthropic's broader effort using an unreleased Claude model to discover and chain browser and OS vulnerabilities into working exploits.

_Security teams weighing AI-assisted patch prioritization can keep up with findings like this on daily.dev._

### Does AI-assisted vulnerability discovery mean attacks are easy to pull off in the real world?

Not necessarily. Black Duck's Andrew Bolster notes exploit success drops sharply once ASLR, stack canaries, and sandboxing are enabled, and VulnCheck found only 1.3% of AI-discovered vulnerabilities are confirmed exploited in the wild. Suzu Labs' Jacob Krell adds that getting a proof of concept to work reliably against an unseen server remains the expensive, hard part, even though AI has made finding bugs cheap.

_Anyone assessing real-world AI exploit risk can track expert takes like these on daily.dev._

## Similar posts on daily.dev

- [Claude Opus wrote a Chrome exploit for $2,283](https://daily.dev/posts/claude-opus-wrote-a-chrome-exploit-for-2-283-vrq54wsru) · The Register · 1 upvotes · 0 comments
- [Claude Code Used to Find Remotely Exploitable Linux Kernel Vulnerability Hidden for 23 Years](https://daily.dev/posts/claude-code-used-to-find-remotely-exploitable-linux-kernel-vulnerability-hidden-for-23-years-vni2dtvdh) · InfoQ · 0 upvotes · 0 comments

---

Tags: [#security](https://daily.dev/tags/security), [#claude](https://daily.dev/tags/claude), [#anthropic](https://daily.dev/tags/anthropic)

[View this post on daily.dev](https://daily.dev/posts/claude-builds-exploits-across-major-browsers-oses-hqofelevw)

```json
{"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]}
{"@context":"https://schema.org","@type":"TechArticle","headline":"Claude Builds Exploits Across Major Browsers \u0026 OSes","url":"https://daily.dev/posts/claude-builds-exploits-across-major-browsers-oses-hqofelevw","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/claude-builds-exploits-across-major-browsers-oses-hqofelevw"},"datePublished":"2026-10-29T18:26:35.609Z","dateModified":"2026-10-29T18:27:08.167Z","description":"An unreleased version of Claude found and chained vulnerabilities into five working exploits against Chrome, Safari, Android, Windows, and Linux. One attack...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/a46728952d182d1390519378a5777bec?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/a46728952d182d1390519378a5777bec?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Dark Reading","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Dark Reading","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/dr","url":"https://daily.dev/sources/dr"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/claude-builds-exploits-across-major-browsers-oses-hqofelevw","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"security,claude,anthropic","timeRequired":"PT6M"}
{"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Dark Reading","item":"https://daily.dev/sources/dr"},{"@type":"ListItem","position":3,"name":"Claude Builds Exploits Across Major Browsers \u0026 OSes"}]}
{"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/claude-builds-exploits-across-major-browsers-oses-hqofelevw#faq","mainEntity":[{"@type":"Question","name":"What vulnerabilities did Claude find and exploit across Chrome, Safari, Android, Windows, and Linux?","acceptedAnswer":{"@type":"Answer","text":"An unreleased Claude model generated five working exploits: a one-click Chrome RCE chaining four bugs in the JS engine, memory protections, process isolation, and federated sign-in; a Safari same-origin-policy bypass chaining four WebKit flaws; an Android exploit extracting data from a locked Pixel 9 via USB-handling bugs; a Windows kernel HTTP/3 denial-of-service; and a zero-click Linux kernel networking RCE with root privileges, all now patched. Developers tracking AI-driven vulnerability discovery can follow coverage like this on daily.dev."}},{"@type":"Question","name":"How many vulnerabilities did Anthropic report to Google's Chrome team using Claude?","acceptedAnswer":{"@type":"Answer","text":"Anthropic reported 62 vulnerabilities to Google, all of which received CVE identifiers, were rated high or critical severity, and were patched by Chrome version 156. This was part of Anthropic's broader effort using an unreleased Claude model to discover and chain browser and OS vulnerabilities into working exploits. Security teams weighing AI-assisted patch prioritization can keep up with findings like this on daily.dev."}},{"@type":"Question","name":"Does AI-assisted vulnerability discovery mean attacks are easy to pull off in the real world?","acceptedAnswer":{"@type":"Answer","text":"Not necessarily. Black Duck's Andrew Bolster notes exploit success drops sharply once ASLR, stack canaries, and sandboxing are enabled, and VulnCheck found only 1.3% of AI-discovered vulnerabilities are confirmed exploited in the wild. Suzu Labs' Jacob Krell adds that getting a proof of concept to work reliably against an unseen server remains the expensive, hard part, even though AI has made finding bugs cheap. Anyone assessing real-world AI exploit risk can track expert takes like these on daily.dev."}}]}
```

