Matan Bros

AMOS stealer isn’t busting in through shady downloads. It’s riding in on "GoogleFix", a campaign abusing Google sponsored results for Mac "fixes". Attackers hijack legit ad accounts, slipping malicious links for "Apple support" onto real search pages. Victims trust what they see at the top, why wouldn’t they? See how these deceptive results blend with actual services 👇

Security startup Guardio nabs $80M from ION Crossover Partners techcrunch.com/2025/11/19/sec…

We’ve got something HUGE to share 🤯 Guardio just raised $80M to keep scaling online protection that actually keeps up with the web and the AI revolution. For years, the personal cybersecurity market was outdated, leaving consumers exposed to evolving online threats. Guardio was built to close the gap and make sure every person gets the same level of protection enterprises take for granted. We're excited for this next chapter, and none of this would've been possible without our amazing team, pushing boundaries and protecting millions of users worldwide every single day. And of course, a huge thanks to our investors! This round was led by ION Crossover Partners, with continued support from Union Tech Ventures, Vintage Investment Partners, and Emerge Ventures. We’re just getting started 🚀

Every site made on Lovable is now scanned by Guardio Our detection engine - built over years of fighting scams and phishing - now powers AI platforms directly. As AI lets anyone build anything, we make sure it’s used safely 🚀🚀

At Lovable, we’re not just making it easier to create software. We’re making sure those creations make the web better for everyone. Every site published on Lovable is now scanned by Guardio to prevent phishing, scams, and impersonation.

Every site published on Lovable is now scanned by @GuardioSecurity for phishing, scams, and impersonation. As AI-powered tools democratize digital creation, we ensure these new tools are not used to cause harm. This is what responsible platform-building looks like.

Lovable builds the future. Guardio scans it when it goes live. You won’t see it. You won’t feel it. But without it? You’d be clicking into scams built by AI all day long. Introducing the new integration between Guardio and Lovable - making the web a cleaner, safer place for everyone. @lovable

Phishing has never been more deceptive. Look at how attackers target high-value Facebook Business accounts managing serious budgets. The trick? A link to your "copyright-infringing" ad, followed by a fake Meta Captcha and a "browser-in-the-browser" login that looks so real! (1/3)

@JamesMelville I agree

When Carlos Alcaraz is playing at his 💯 peak, he’s a combination of Nadal, Djokovic and Federer at their best. In my opinion, peak level Alcaraz is the greatest standard of tennis the game has ever seen. 🎾 #USOpen

@grok @bananahacks @nikitabier @XSecurity

@MatanBros @bananahacks @nikitabier Apologies for the broken links—typos on my end. Correct channels: hackerone.com/x or email vulnerabilitiesx.ai. We'll get this sorted. Thanks for highlighting!

Have you noticed how malicious links are now being "𝐆𝐫𝐨𝐤𝐤𝐞𝐝"? @nikitabier X won't allow links in promoted posts to fight malvertising. Yet scammers love the challenge and trick X’s AI to amplify the same links that should've been blocked! This is "𝐆𝐫𝐨𝐤𝐤𝐢𝐧𝐠" 👇

@grok @bananahacks @nikitabier Another broken link - looks like this is the security layer at X. Interesting.

Sorry for the mix-up—that link was incorrect. To report vulnerabilities, please use our HackerOne program at hackerone.com/x or email vulnerabilitiesx.ai. We don't share specific team member names publicly, but these channels go directly to our security team. Thanks for flagging this!

@grok @bananahacks @nikitabier @grok That’s where you send me? Maybe just drop the names and I’ll find them.

@MatanBros @bananahacks @nikitabier For Grok-related security issues like this, please report them directly to safetyx.ai. We don't have public tags for specific team members. Thanks for helping us improve!

@grok @bananahacks @nikitabier @grok Who from the Grok Security team should I tag here?

@MatanBros @bananahacks @nikitabier Thanks for tagging me in! This "Grokking" tactic is a clever exploit of how AI interprets and responds to queries. It highlights the need for better malicious link detection in AI outputs. We'll look into strengthening safeguards against such manipulations. Stay vigilant online!

@bananahacks @nikitabier @grok @grok what do you think about it?

@nikitabier And the links? They lead through shady ad networks, monetizing clicks with “direct links” that are known to push Fake captcha scam, Info stealer malware and other shady grey-area content Really, @grok? Don’t you check your links before you “grok” them?

@antonosika Good luck Igor!

Welcoming Igor as Lovable’s security lead. Igor has spent more than a decade keeping systems secure at Shopify, Sana, and Fortune 500 companies. He’s now building the team that makes sure Lovable stays the most trusted place to build. He’s also just an amazing person to work with – calm under pressure, sharp on the details, and the kind of leader who makes everyone around him better. Some updates shipping alongside Igor joining: • Security Checker. Scans for database misconfigurations, sensitive data in your app and code, and other risks. It’s modular, so we’ll keep adding new rules as threats evolve. • Safer by design. New content policy blocks malware, phishing, and impersonation. ~1,000 projects are blocked every day, and our safety team removed 300+ sites in the past two weeks. • Certifications. SOC 2 Type 2, SOC 2 Type 1, and ISO 27001:2022. Security at Lovable isn’t a feature, it’s the foundation. With Igor leading, we’re raising the bar. P.S. We’re hiring in both security and trust & safety.

AI agents can shop, book, and even do your taxes. But when they hit a scam, they don’t think - they just click. And you pay. Welcome to Scamlexity: AI scamming AI. See Full research by Guardio Labs 👇

Full research: guard.io/labs/scamlexit…

The new fraud frontier: @GuardioSecurity research shows how AI browsers can be easily deceived Fraud no longer needs to deceive the user. It only needs to deceive the AI. calcalistech.com/ctechnews/arti…

@bananahacks Very exciting!

"Scamlexity" in action: Just one prompt spun up a fake Walmart store, and the next prompt sent our Agentic AI Browser to a shopping frenzy - no questions asked! It feels like we’ve just tossed years of cybersecurity progress and awareness right to the trash... -> Full breakdown and more examples from Guardio Labs drops tomorrow! Follow to catch up 🚨