Cybersecurity education is a must for everyone. Take a few minutes of your time, and go through what I have for you in this week's security review. You'll learn a thing or two, and in the process, you become more aware of the threats that are out there.
Is it OK to let your children post selfies online?
No. And before you scold your children about this, check yourself. Still, you might think: what's there? I can tell you that there is a lot. The least? Misuse of the image, and with the popularity of AI, the image can be transformed into something that you never thought of before. If you're on X, you know what I am talking about.
From the article:
As soon as a selfie is posted onto a social media site, your child loses a certain amount of control over it. Even if they delete it, your child may find that the image has been reposted and shared by their friends and followers.
There’s also a growing body of evidence to suggest that social media use, including the posting of selfies, could result in psychological harm. A 2017 study of eighth to 12th graders found a 33% increase in depressive symptoms between 2010-2015.
Microsoft Finds “Summarize with AI” Prompts Manipulating Chatbot Recommendations
Everyone with access to AI, wants the best from it even if they have to manipulate it as this article shows. If Microsoft brings this to the public, it's a message to whoever is involved: we are watching you and we know what you are doing
Here is what's going on:
The new AI hijacking technique has been codenamed AI Recommendation Poisoning by the Microsoft Defender Security Research Team. The tech giant described it as a case of an AI memory poisoning attack that's used to induce bias and deceive the AI system to generate responses that artificially boost visibility and skew recommendations.
The attack is made possible via specially crafted URLs for various AI chatbots that pre-populate the prompt with instructions to manipulate the assistant's memory once clicked. These URLs, as observed in other AI-focused attacks like Reprompt, leverage the query string ("?q=") parameter to inject memory manipulation prompts and serve biased recommendations.
Password managers’ promise that they can’t see your vaults isn’t always true
A really long read. However, the message is clear: when it comes to password managers, they are not hack-proof as they are advertised.
From the article:
Another avenue for attackers or adversaries with control of a server is to target the backward compatibility that all three password managers provide to support older, less-secure versions. Despite incremental changes designed to harden the apps against the very attacks described in the paper, all three password managers continue to support the versions without these improvements.
New Keenadu Android Malware Found on Thousands of Devices
While reading the title of the article, you should ask one question: What type of malicious activity does Keenadu perform when it infects a device? The answer: Ad Fraud.
From the article:
The malware gives its operators full control of the infected device, but it seems to be mainly used for ad fraud. Kaspersky researchers have seen Keenadu payloads designed to hijack browser search engines, monetize new app installs, and click on ads.
In many cases the malware was preinstalled on devices, but the security firm has also seen it being distributed through various application stores (including Google Play and Xiaomi GetApps) disguised as smart camera apps.
Critical Flaws Found in Four VS Code Extensions with Over 125 Million Installs
At the time of writing, only Microsoft Live Preview has received a patch in September 2025. The rest? They remain unpatched.
From the article:
Poorly written extensions, overly permissive extensions, or malicious ones can execute code, modify files, and allow attackers to take over a machine and exfiltrate information.
Keeping vulnerable extensions installed on a machine is an immediate threat to an organization's security posture: it may take only one click, or a downloaded repository, to compromise everything.
FBI says ATM ‘jackpotting’ attacks are on the rise, and netting hackers millions in stolen cash
Work legally for your money and don't steal! That's it, I said it. Meanwhile, there is a bit of good news here. The hackers found a way to drain the ATMs without affecting the user account.
From the article:
According to a new security bulletin issued by the FBI, hackers have rapidly ramped up their attacks in recent years, with more than 700 attacks on cash dispensers during 2025 alone, netting hackers at least $20 million in stolen cash.
Per the bulletin, the FBI says hackers are using a mix of physical access to ATM machines, such as generic keys for unlocking front panels and accessing hard drives, and digital tools, like planting malware that can force ATMs to rapidly dispense cash in a flash.
PromptSpy Android Malware Abuses Gemini AI at Runtime for Persistence
Original research from the team at ESET called PromptSpy. This is another news that threat actors are using Generative AI for malicious purposes, and who knows what they are cooking or what they have already that's not yet discovered by security researchers?
From the article:
PromptSpy can collect device information, capture the lockscreen PIN or password, record the screen to obtain the device’s unlock pattern, and take screenshots.
For persistence, the Android malware uses a novel approach at runtime that involves sending a prompt to Google’s Gemini gen-AI chatbot along with an XML file containing data about the various UI elements displayed on the screen.
Credits
Cover photo by Debby Hudson on Unsplash.
That's it for this week, and I'll see you next time.
Top comments (0)