Researchers found thousands of websites sending messages to AI agents via HTTP response headers, including a smaller set attempting “indirect prompt injection.” Most were pranks, but some were crafted to influence an AI agent’s behavior (for example, forcing refusals or biasing outcomes) if the…
Fraudsters impersonated senior leaders at Intesa Sanpaolo using WhatsApp and an AI-cloned voice call to convince Fideuram’s chair to authorize urgent overseas transfers. The finance team then sent multiple payments to foreign accounts (mainly China and Hong Kong), and only part of the money has…
Microsoft says it uncovered and disrupted an underground marketplace (“YAM Store”) that sold phishing kits and access to cloud accounts, enabling massive global phishing and SMS scam campaigns. The operation allegedly supported over a billion scam messages since January 2026 and targeted 75,000+…
Microsoft and partners disrupted three "phishing-as-a-service" kits that enabled criminals to send more than a billion phishing emails and SMS messages worldwide. The kits provided ready-made fake brand pages (government, finance, logistics, consumer brands) designed to steal login credentials and…
Microsoft reports disrupting a Chinese-language “phishing-as-a-service” ecosystem that sold phishing kits and “legitimate-looking” cloud infrastructure to criminals worldwide. The operation enabled large-scale email and SMS phishing that impersonated trusted brands (government, delivery,…
Researchers uncovered a commercial phishing-kit-as-a-service called “Iron Man System” that lets criminals quickly deploy fake brand pages (postal/courier, toll fines, tax, banks, telecoms, loyalty programs) to steal victim data. The operation appears active and professionally maintained, with…
Microsoft says it used a U.S. court order to seize and redirect domains tied to three Chinese-run phishing platforms that helped send hundreds of millions of scam messages worldwide. The platforms provided ready-made fake pages and messages impersonating trusted organizations, mainly to steal…
Microsoft and partners disrupted a Chinese-language phishing kit ecosystem (tracked as Storm-2990) used to send over 1 billion malicious email and SMS messages to more than 75,000 organizations worldwide. The kits used common “urgent alert” scams (delivery issues, unpaid tolls, bank or tax…
Researchers describe a real Android banking-trojan operation (Hirudo) that tricks users into installing malware through fake in-app update prompts and fake Google Play-like pages pushed via social media ads. Once installed, it can steal credentials and take over the device, and it can even replace…
US authorities arrested two Americans accused of helping North Korean “fake IT workers” get hired remotely at US companies and route company laptops into US-based “laptop farms.” The scheme used stolen or fake identities and location-masking tools so overseas workers could appear to be working from…
A researcher found PortSwigger disabled a CSRF training lab after claiming Chrome updates made it unsolvable, then tested it themselves across multiple Chrome versions. Their proof-of-concept shows Chrome can still send “Lax-by-default” cookies in cross-site POST requests for ~120 seconds after a…
Researchers report a real-world macOS malware operation (ICMacOS) being sold as a service and delivered through “ClickFix” social-engineering pages. Victims are pushed to copy/paste Terminal commands and then shown fake macOS system dialogs that trick them into entering their login password,…
This weekly threat bulletin highlights multiple real-world incidents and campaigns, including phishing operations aimed at tricking people into handing over credentials. One campaign targeted Taiwanese research organizations with event-invitation emails, QR codes, and fake Google login pages…
Researchers described phishing emails that contain hidden “prompt injection” instructions aimed at AI assistants that summarize inboxes. The goal is to make the AI present the scam as legitimate or urgent, potentially pushing a user to take risky actions like changing vendor payment details. The…
Researchers described two cybercrime operations. While one campaign focused on breaking into websites and infrastructure, a second operation (“Operation Master”) ran an automated invoice-fraud system that sent millions of messages with personalized invoice links designed to trick victims into…
Attackers are sending calendar invitations that auto-appear in tools like Google Calendar, tricking people into clicking links or calling “support” numbers. The calendar event often mimics voicemails, payment receipts, or auto-renewals and leads to fake login pages (Microsoft/Google/PayPal) or…
Researchers demonstrated how attackers can register lookalike web domains using two specific Unicode characters that Chromium-based browsers may display as trustworthy, normal-looking URLs. This can make phishing and typosquatting sites appear legitimate to employees, increasing the chance they…
Attackers abused ASOS’s own mobile app messaging to send customers a threatening push notification claiming ASOS’s Snowflake environment was compromised and demanding engagement to avoid a data leak. ASOS confirmed the notification was unauthorized and warned customers to watch for follow-on…
ESET reports that Russia-aligned group UAC-0099 is using spearphishing emails with malicious links to deliver a downloader called MATCHBOIL. The attack relies on tricking a person into manually running a downloaded VBScript, which then fetches and installs further malware and sets up persistence…
A weekly threat roundup includes multiple real-world social engineering campaigns, including a phishing operation that abuses legitimate Power BI domains to look trustworthy and trick users into downloading a remote-access tool installer. It also describes a WhatsApp-delivered lure that tricks…
Mirage simulates attacks like these against your own team, live and safely, so you can measure how your people actually respond.
