Danielle Aminov retweeted
I guess you could say the past 48 hours have been a bit busy for cyber defenders
Attackers are using AI to make Entra ID registered devices look… normal.
Defenders must adapt:
• Drop static identifiers
• Focus on behavioral detection
• Correlate activity across the attack chain
wiz.io/blog/detecting-entra-…
Our AI attacker autonomously made its way into @Snowflake's internal Jira, accessing sensitive data simply by opening a public GitHub issue with a crafted title.
The flaw was introduced 5 days earlier by "Copilot Autofix powered by AI" 🪄 🧵
Danielle Aminov retweeted
While the keyv worm's tail is still unfurling, it's important to recognize that not all malicious packages are alike. In this case, the vast majority of the exposure is so far isolated to the initial seed compromise:
Danielle Aminov retweeted
A lot of cloud service providers beyond AWS allow you to store data in something that looks and acts like S3, but there are differences worth investigating. Check out my latest article discussing the security risks that carry over and the assumptions that break.
wiz.io/blog/s3-clones-in-the…
Danielle Aminov retweeted
This might be unfair nitpicking, but I'm bothered by the fact that no one involved in writing / reviewing Anthropic's blogpost thought to use the well-established technical term for the technique described in this paragraph (dependency confusion).
Replying to @AminovDanielle
@AminovDanielle and i did the one thing harder than dealing with supply chain attacks - which is try to be cool on camera while talking about them
Check it out! 👩🏻🐰👩🏽💅
🏆 Introducing Wiz Brackets, where the biggest cyber threats go head-to-head. 🥊
Our threat researchers @AminovDanielle & @merav_br break down which attack wins on complexity, impact, novelty & responder pain
Watch the full episode 🎬
youtu.be/fFDKDLRQneQ
🏆 Introducing Wiz Brackets, where the biggest cyber threats go head-to-head. 🥊
Our threat researchers @AminovDanielle & @merav_br break down which attack wins on complexity, impact, novelty & responder pain
Watch the full episode 🎬
youtu.be/fFDKDLRQneQ
Danielle Aminov retweeted
MCP servers can unintentionally expose sensitive data or allow anonymous users to perform destructive actions on internal databases - we scanned a wide variety of servers and found plenty of evidence that this risk isn't theoretical:
wiz.io/blog/the-risk-hiding-…
Danielle Aminov retweeted
We (+@liormama) gained full access to EVERY database on Azure's flagship DB service 🤯
And all we needed was a single key 🔑
This is the story of #CosmosEscape 🧵👇
Danielle Aminov retweeted
Sharing our observations so far in terms of wp2shell exploitation in the wild, by @shahardorf @GiliTikochinki
wiz.io/blog/wp2shell-cve-202…
Danielle Aminov retweeted
Way back in December 2025 @AmitaiCo & I hyped GitHub as a threat intelligence source on the @BlackHatEvents Europe stage.
I think we were immediately vindicated!
Despite all that has happened in the GitHub threat landscape since, it's still a masterclass in GH TI: youtube.com/watch?v=JZUV8dY7…
🎙️ NEW PODCAST: AI just uncovered a #Linux bug hiding since 2017 👀
Eden & @AmitaiCo sit with @tjbecker & Jacob Newman from @xint_official to unpack CopyFail, the privilege escalation flaw their autonomous AI agents discovered affecting Linux machines worldwide.
🚨 Meet "Dirty Frag": a new Linux kernel privilege escalation, fresh off the heels of Copy Fail.
Wiz Research is tracking CVE-2026-43284 and CVE-2026-43500 (also known as Copy Fail 2), discovered by Hyunwoo Kim (@v4bel). No official patches are out yet, and a public PoC already exists.
🔑 The technical bit:
A vulnerability chain in the ESP (IPsec) and RxRPC kernel subsystems, with two deterministic page-cache write primitives in the same spirit as Dirty Pipe. Ubuntu, RHEL 8/9/10, AlmaLinux, Fedora, openSUSE, and more are all affected.
🛡️ Slightly less risky for containers:
Exploitation usually requires CAP_NET_ADMIN, so hardened Kubernetes setups with default seccomp profiles are at lower risk. VMs and less restricted environments should still take this seriously.
🔧 What to do now:
Until patches drop, you can temporarily disable the vulnerable kernel modules (esp4, esp6, rxrpc). Full mitigation and detection guidance in the blog.
Wiz customers: pre-built queries and a live advisory are already in the Threat Intel Center. We'll keep updating both as more info comes to light.
Full research from Merav Bar and @ramimacisabird: wiz.io/blog/dirty-frag-linux…
Danielle Aminov retweeted
npm package "intercom-client" version 7.0.4 is compromised (still available to download as of now), likely related to Mini Shai Hulud campaign by TeamPCP based on shared characteristics
🎉 IT'S OFFICIAL: @wiz_io joins @Google to secure the AI era.
This is a massive moment for our customers and our team.
Thank you to every customer, partner, and Wizard who made this moment possible 💙
We can't wait to share what's next.
wiz.io/blog/google-closes-de…
Danielle Aminov retweeted
Replying to @shahardorf
@shahardorf & I found a phishing campaign abusing oauth applications in Entra in more than 50 organizations! And i promise you that in this blog we explain how you can do it too! And provide all the IOCs 🤭 It's one of these blogs i would enjoy reading!
wiz.io/blog/detecting-malici…
Introducing my Bug Bounty Masterclass. 100% free.
I've made $2,000,000+ finding security bugs. I spent the last year turning my methodology into a complete blueprint.
4 hours of video - foundations, reconnaissance, web proxies, hands-on challenges, and certification.
Finish it in a weekend and start hacking real-world applications 🐞


















