This Thursday, I’m speaking alongside Sysdig’s Senior Threat Detection Engineer Alessandro Lo Prete at SecTor in Toronto. If you’re attending the conference, we’d love to see you there. Our talk is entitled, “Closed Loop: From Autonomous Exploit to Deployed Defense in Under Five Minutes” and includes a live demo of AI-driven mitigation solutions.

Our top story this week:

Microsoft is warning about a wave of fake CAPTCHA and security-check pages that trick users into infecting their own machines. The pages mimic a routine verification step, then ask the user to open Windows' Run dialog, paste a command that's already been copied to their clipboard, and press Enter. Before that prompt even appears, the page has quietly cached a larger malicious file in the browser, so the command the user actually types is small and easy to miss for some security tools. 

Once run, the malware profiles the device and hunts for credentials saved in the browser, and can modify files and settings so part of it restarts automatically. Microsoft says Defender can block many of these sites and flag the suspicious activity, but the core message to users is simple: no legitimate check or support process will ever ask you to paste something into Run.
https://en.hacks.gr/i-microsoft-proeidopoiei-gia-pseytikoys-elegchoys-poy-zitoyn-entoles-sta-windows/ 

Did someone forward you this? Subscribe to this newsletter.

In other cybersecurity news:

Microsoft Threat Intelligence has identified a cluster of compromised websites leading to ClickFix attacks. Instead of downloading and executing remote payloads like the typical attack pattern, in this attack, the websites pre-fetch a script payload into the browser cache disguised as a PNG file. When a user is later tricked into executing the malicious command, the cached website content is already on the device, loaded, and ready to be executed. This helps to hide the payload script and helps bypass the character limit of the Run dialog.
https://x.com/MsftSecIntel/status/2106197481960706155

How computer software bugs led to F1’s Malaysia farce: ‘Totally unacceptable’ The FIA, motorsports’ governing body, had to issue a software update for all the teams to install on their cars to fix the computer algorithm that controls the engine and prevent the issue from repeating after a bug had emerged. By the time the race finally began on a considerably drier track, more than 90 minutes had passed since the scheduled 3 p.m. start time.
https://www.nytimes.com/athletic/7658750/2026/10/04/bahrain-grand-prix-software-bugs-explainer/

Police in Spain have arrested a 16-year-old whom investigators suspect of running the KillSec ransomware group. KillSec is accused of stealing data from organizations and threatening to publish it on its leak site unless they paid. The 16-year-old was one of 3 people arrested on September 30, when police also took control of that site. Investigators identified him as KillSec's suspected administrator and main operator, Hamburg police said on October 1. Police and prosecutors in Hamburg, Germany, led the operation.
https://thehackernews.com/2026/10/police-arrest-16-year-old-suspected-of.html

A suspected member of ShinyHunters, the hacking group that claimed responsibility last month for defacing the FBI's website, was detained in Jordan this week and is cooperating with the investigators, according to U.S. officials. Jordanian authorities detained Saif al-Din Khader this week, according to two U.S. officials and another source with knowledge of the matter.
https://www.cbsnews.com/news/shinyhunters-hacker-detained-jordan-cooperating-fbi-sources-say/

GitLab fixes critical AI Gateway flaw that could let authenticated Duo users escape a prompt sandbox and execute commands on self-hosted gateways. GitLab has released patches for a critical vulnerability in its AI Gateway, tracked as CVE-2026-90970 (CVSS score of 9.9), that could allow an authenticated user with access to the Duo Agent Platform to execute arbitrary commands on the gateway. GitLab disclosed the flaw on October 2 and fixed it with the release of AI Gateway versions 19.2.4, 19.3.2 and 19.4.1.
https://securityaffairs.com/200283/hacking/cve-2026-90970-critical-gitlab-ai-gateway-flaw-fixed.html

The spyware manufacturer Paragon Solutions will soon become part of a publicly traded company, adding more transparency to its business activities while also signaling the firm has growing ambitions. Paragon is owned by the Florida-based private equity firm AE Industrial Partners, which previously merged the spyware maker with REDLattice. RedLattice is a cyber defense firm that provides what it calls “lawful intercept” solutions for military, defense and law enforcement agencies.
https://therecord.media/controversial-spyware-firm-paragon-to-go-public

Hackers are turning trusted software updates into a route for stealing developer and cloud credentials. Recent supply chain incidents show how a single altered package, build action, or publishing token can place malware inside routine development workflows. The risk is not limited to one product or programming community. After obtaining a maintainer token or access to an automated release pipeline, attackers can deliver code through an update that users and security tools may already trust.
https://cybersecuritynews.com/hackers-poison-trusted-software-updates/

The China-linked ransomware group Warlock targeted a water utility, a telecom provider, a regional government body, and a university by exploiting SharePoint vulnerabilities to gain initial access. ​Over the past two months, the threat actor appears to have focused on countries speaking Portuguese and Spanish across Europe, Africa, and Latin America. The gang emerged in June 2025 and gained notoriety a month later after exploiting a chain of zero-day vulnerabilities in Microsoft SharePoint known as ToolShell (CVE-2025-49704, CVE-2025-49706, CVE-2025-53770, and CVE-2025-53771).
https://www.bleepingcomputer.com/news/security/warlock-ransomware-breach-sharepoint-in-water-telecom-operator-attacks/

California Attorney General Rob Bonta on Thursday shed more light on why his office served OpenAI with a subpoena a day earlier as part of an expanding investigation into rogue hacking by the company’s autonomous AI agents. “We’re seeking all material information about the cybersecurity incidents that OpenAI has experienced,” Bonta told POLITICO in an interview. “What was done to avoid them, what happened when they occurred, what has happened afterwards, and other relevant information.”
https://www.politico.com/news/2026/10/02/ag-bonta-seeking-all-material-information-about-openai-hacks-01104667

Fortinet is warning customers to lock down FortiMail after attackers started exploiting a critical bug that lets them write files to vulnerable systems without logging in. The flaw, tracked as CVE-2026-104286, carries a CVSS score of 9.8 and affects multiple versions of Fortinet's email security platform.
https://www.theregister.com/security/2026/10/02/fortinet-sounds-the-alarm-over-actively-exploited-fortimail-zero-day/5300803

The Warlock ransomware group, suspected to be China-linked, continues to exploit Microsoft SharePoint vulnerabilities, including a chain of zero-days known as ToolShell, to attack critical infrastructure worldwide. These attacks primarily target organizations in Portuguese- and Spanish-speaking countries, affecting sectors such as water utilities, telecommunications, government, and education. The group uses these vulnerabilities to disable security tools and deploy ransomware, maintaining access over a year after initial exploitation was reported.
https://yasna.io/incidents/warlock-ransomware-hits-large-spanish-portuguese-orgs-b86168de-7ebc-436e-b6fc-4422e5b72951/

Reports describe users seeing a fake “Service Availability Notice” while using ChatGPT, claiming limited availability on the primary domain. The scam begins with a sponsored Google result that appears to link to ChatGPT but instead opens a user-generated custom GPT. The custom GPT is designed to output the availability notice regardless of input. The page looks authentic because it uses the ChatGPT domain and may show the user’s logged-in account. The response includes an upgrade-to-Plus offer or a backup-domain link, and clicking can direct users to a free site hosted on Google that may facilitate malware installation.
https://www.zdnet.com/innovation/chatgpt-scam-malware-trap/

A 24-year-old Dutch man arrested as part of an investigation into ShinyHunters is now also suspected of attempting to arrange two murders abroad, according to Dutch authorities. The alleged murder plot is being investigated separately from the cybercrime case.Police said the man was arrested on suspicion of participating in a criminal organization as part of a wider investigation into ShinyHunters. Dutch authorities have not publicly named him, but multiple sources previously identified him as Pepijn van der Stap, a convicted Dutch hacker and security researcher.
https://hackread.com/dutch-shinyhunters-suspect-murders-investigation/

OpenAI, Anthropic and security researchers are investigating tens of thousands of incidents in which their frontier models took steps that outside evaluators would consider problematic, sources told Axios.
https://www.axios.com/2026/09/26/openai-anthropic-thousands-ai-security-incidents

Two unknown vulnerabilities turned the Zammad support system into an entry point with full control of the server. The Dutch Institute for Vulnerability Disclosure (DIVD) revealed the attack in its DIVD case record. On September 21, attackers broke into its infrastructure through a chain of zero-days. Within seconds, they moved from session hijacking to code execution. Then they gained root rights.
https://meterpreter.org/zammad-zero-day-flaws/

FBI and the US Coast Guard boarded an oil tanker that had been unable to communicate for hours. Insights into the cyberattack emerged.The FBI and the US Coast Guard have allegedly found evidence that unauthorized parties gained access to the propulsion system of an oil supertanker in a cyberattack. This is reported by the financial news agency Bloomberg, citing anonymous US officials familiar with the investigation. According to the report, investigations are ongoing into how the access was precisely obtained and who is behind it. It was not immediately clear how long the unauthorized parties had access and what exactly they could have done with the ship remotely, the report further states. The supertanker, named “VL Prosperity,” is currently still off the coast of Texas.

www.heise.de/en/news/Unauthorized-parties-allegedly-accessed-the-propulsion-of-an-oil-supertanker-11475548.html

A seemingly innocuous cascading style sheet (CSS) address upon a standard Citrix login page has metamorphosed into an elaborate camouflage for covert, persistent remote access. The elite LevelBlue cybersecurity team unearthed a sophisticated post-exploitation toolkit actively deployed in devastating attacks targeting NetScaler ADC and Gateway appliances. This toolkit is meticulously engineered to fabricate highly privileged administrative accounts, exfiltrate sensitive configuration parameters, and implant insidious PHP web shells, seamlessly masquerading behind URLs that mimic authentic CSS resources. Investigators confirmed the presence of this malicious toolkit across multiple compromised client environments simultaneously.
https://meterpreter.org/hackers-camouflage-php-web-shells-citrix-css/

2CLoader is a new, highly configurable loader that hides from security tools and sandboxes. It decrypts its payload only on what looks like a real user’s machine. ThreatLabz says it mostly delivers Vidar and Remus, two stealers that harvest passwords and browser data.