@SecurityMB

Improving the world’s security at Google. Opinions are mine.

 Zurich, Switzerland
bentkowski.info/
Joined September 2014
  • Tweets1.2K
  • Following284
  • Followers11.3K
87Media
We just published two blog posts about PageBreak, Google's AI web security scanner (that I've helped develop for the better part of the year). We share our approach and a couple of findings, enjoy! - blog.google/security/agentic… - bughunters.google.com/blog/p…

Agentic Hacks, Real Proofs: Inside Google's PageBreak Project

Security landscape in 2026The application of Large Language Models (LLMs) to security scanning has revolutionized the vulnerability management landscape. But, it has als…

blog.google
12
56
5
316
32,164
“And no, AI isn't perfect. But Christ, anybody who points to the problems at AI had better be looking in the mirror and pointing at themselves at the same time. Because it's not like natural intelligence is always all that great either.” lore.kernel.org/linux-media/…
2
1
13
2,304
This is a really awesome summary of how to approach bypassing HTML sanitizers, I highly recommend to check this out!
Masato Kinugawa@kinugawamasato
Jun 12
DOMPurifyバイパスの歴史をまとめたページ。HTMLサニタイザーバイパスのノウハウがぎっしり詰まってる。 特に今年に入ってから発見された<selectedcontent>のバイパスは驚いた github.com/cure53/DOMPurify/…

Attack Classes & Bypass History

DOMPurify - a DOM-only, super-fast, uber-tolerant XSS sanitizer for HTML, MathML and SVG. DOMPurify works with a secure default, but offers a lot of configurability and hooks. Demo: - cure53/DOMPurify

github.com
1
12
2,147
Have been working on agentic vulnerability discovery for a while now, happy to share about CodeMender! 🚀🚀
Google
@Google
May 19
While AI is helping us tackle big challenges, it also poses new potential security risks. That’s why we’re focused on building tools that keep the broader ecosystem safe — like CodeMender, a code security agent that automatically finds and fixes critical software vulnerabilities. Today, we’re inviting a select group of experts to test a new CodeMender API, and we’ll be launching it more broadly soon. #GoogleIO
Text reads "CodeMender" over an abstract pink and blue grid illustration.

ALT Text reads "CodeMender" over an abstract pink and blue grid illustration.

1
4
58
7,760
Anyone I know interested in joining the Google Security Team in Zurich? Let me know, I can give a referral :D Here's the job posting: google.com/about/careers/app…
11
35
2
222
29,978
That must be the worst captcha I’ve ever seen.
1,367
That was a nice bug, thanks for the shoutout!
Intigriti
@intigriti
Jan 23
Replying to @intigriti
2️⃣ XSS in GMail's AMP4Email via DOM Clobbering Michał Bentkowski (@SecurityMB) exploited DOM clobbering to achieve XSS in Gmail's AMP4Email feature. Found that AMP4Email allowed id attributes, which could be leveraged to overwrite JavaScript variables and bypass Google's strict security filters. research.securitum.com/xss-i…
1
1
11
2,276
Why the "Agents Rule of Two" is flawed shhnjk.substack.com/p/why-ag…
1
3
13
2,346
Cross-Site ETag Length Leak blog.arkark.dev/2025/12/26/e… I just posted the author writeup for impossible-leak in SECCON CTF 14 Quals. As far as I know, this is a new XS-Leak technique! The ETag header can become a side channel :)

Cross-Site ETag Length Leak | XS-Spin Blog

A novel XS-Leak technique that turns ETag length differences into a cross-site oracle via 431 errors and History API.

blog.arkark.dev
4
99
4
355
44,153
Has there been some email leak from @Canva? Just got a spam message for an email used only for Canva.
2
1
2
2,097
Ah, it's probably from the old leak: haveibeenpwned.com/Breach/Ca…

Have I Been Pwned: Canva Data Breach

In May 2019, the graphic design tool website Canva suffered a data breach that impacted 137 million subscribers. The exposed data included email addresses, usernames, names, cities of residence and...

haveibeenpwned.com
1
1
2
1,012
We launched a redesigned Project Zero website today at projectzero.google ! To mark the occasion, we released some older posts that never quite made it out of drafts. Enjoy!

Google Project Zero

Make zeroday hard

projectzero.google
7
62
2
365
46,547
Interested in the security of AI Agents 💁🛡️? Then you've likely heard of "prompt injection", but do you know what "task injection" is? If you're curious, check out our latest post for a description and some real-world examples we discovered. bughunters.google.com/blog/4… bughunters.google.com/blog/4…

Blog: Task Injection – Exploiting agency of autonomous AI agents

Check this post to find out what a Task Injection attack is, how this type of attack differs from Prompt Injection, and how it is particularly relevant to AI agents designed for a wide range of...

bughunters.google.com
4
67
3
284
32,607
my new blogpost is out!! this one talks about a new web vulnerability class i discovered that allows for complex interactive cross-origin attacks and data exfiltration and i've already used it to get a google docs bounty ^^ have fun <3 lyra.horse/blog/2025/12/svg-…

SVG clickjacking

A novel and powerful twist on an old classic.

lyra.horse
20
76
4
430
40,081
JavaScript’s lexer ambiguity in action. Might fool you and some weak parsers. Demo: jsfiddle.net/yo0a24dj/
8
41
9
208
41,851
We published a blogpost about SafeContentFrame - a library for rendering untrusted content inside an iframe. The library is a big party of what I've been up to in the few last years! Check out the blog and take a slice of my birthday cake 🎂! bughunters.google.com/blog/6…
6
52
202
18,884
Rendering untrusted web content is fraught with security risks 🕸️ 🛡️. Read how SafeContentFrame, a new TypeScript library, offers a robust solution for isolating web content and protecting against threats like XSS and side-channel attacks. goo.gle/3K5DRQJ

Blog: Beyond Sandbox Domains: Rendering Untrusted Web Content with SafeContentFrame

Rendering untrusted web content is fraught with security risks. Learn how SafeContentFrame, a new TypeScript library, offers a robust solution for isolating web content and protecting against threats...

bughunters.google.com
3
18
102
8,443
I'm happy to release a script gadgets wiki inspired by the work of @slekies, @kkotowicz, and @sirdarckcat in their Black Hat USA 2017 talk! 🔥 The goal is to provide quick access to gadgets that help bypass HTML sanitizers and CSPs 👇 gmsgadget.com/ 1/4
12
172
3
457
43,560
It started! pV!
terjanq@terjanq
25 Jun 2025
Google CTF will start in less than 48h from now. Make sure not to miss the great challenges we've prepared this year!! Can't describe how exicted I am for it 😶
3
29
4,916
Google CTF is on! Here's a challenge that I created: capturetheflag.withgoogle.co…. Good luck 😀
26
148
13,280
Load more