Research Saturday
Recent Episodes
Ismael Valenzuela, Vice President of Labs, Threat Research and Intelligence at Arctic Wolf, sits down with Dave to discuss their work tracking Anubis. Arctic Wolf Labs details a series of 2026 Anubis ransomware intrusions, revealing affiliates using stolen VPN credentials and exploiting CitrixBleed 2 to gain initial access. Attackers then blended into legitimate IT activity by deploying RMM tools, using RDP and PsExec for lateral movement, stealing credentials, and establishing tunnels and proxies for persistence and exfiltration. The research highlights a repeatable attack chain defenders can disrupt before encryption, from suspicious remote access and unauthorized RMM deployment to credential theft, security-tool tampering, and ransomware execution.
Today we are joined by Crystal Morin, Senior Cybersecurity Strategist, and Michael Clark, Senior Director of Threat Research, at Sysdig, sharing their work on "LLMjacking evolved: Attackers are using stolen AI compute to build offensive agentic tools." The Sysdig Threat Research Team observed an attacker abusing an exposed, unauthenticated Ollama server as the “brain” for an automated offensive security tool. The AI-powered framework can fingerprint services, identify vulnerabilities, craft exploits, extract credentials, and orchestrate attacks toward command execution, with researchers capturing the tool while it was still under active development. The activity highlights how LLMjacking is evolving from simply stealing AI compute for profit into using stolen model capacity to build increasingly autonomous offensive capabilities.
Today we are joined by Aaron Beardslee, Manager of Threat Research at Securonix, discussing "Analyzing SHEET#CREEP: SHEETCREEP is up again with different config obfuscation." Securonix researchers have identified an evolved version of the SHEETCREEP espionage campaign, using a diplomatic-themed ISO phishing lure to deliver a C# remote access trojan targeting Indian diplomatic interests. The malware abuses the Google Sheets API as a stealthy command-and-control channel, with researchers identifying 91 active victim tabs, including a high-confidence target in Pakistan. The campaign, assessed with moderate confidence as linked to Pakistan-aligned APT36, has added XOR-obfuscated configurations and other anti-analysis techniques to evade detection and maintain persistent access.
TranscriptThe botnet that scouts before it strikes.
Today we are joined by Ian Goldin, Senior Lead Information Security Engineer, and Mike Horka, Principal Information Security Engineer, from Lumen's Black Lotus Labs, discussing their research entitled "Expanded JDY IoT and SOHO botnet enables rapid vulnerability exploitation." Black Lotus Labs has uncovered a major resurgence of the JDY botnet, a China-nexus reconnaissance network now comprising more than 1,500 compromised SOHO and IoT devices. The botnet uses these devices to conduct targeted scanning and fingerprinting, helping threat actors rapidly identify vulnerable infrastructure—sometimes within hours of a new vulnerability disclosure—and appears to have a particular focus on U.S. military-related networks. The research highlights how compromised routers and IoT devices can be turned into distributed reconnaissance infrastructure that evades traditional IP-based defenses and supports follow-on exploitation.
TranscriptA little help from your search engine.
Today we are joined by Brian Hussey, SVP of Howler Cell Threat Services at Cyderes, discussing their work on "Bad Ads, Worse Binaries: Fake Claude Code Installer Drops Infostealer." Howler Cell identified an SEO poisoning campaign targeting people searching for Claude Code installation guides, using a fake Anthropic page and a ClickFix lure to trick victims into running a malicious MSHTA command. The attack uses a six-stage, largely fileless chain that employs an MP3/HTA polyglot, PowerShell obfuscation, AMSI bypasses, per-victim infrastructure, and in-memory execution to evade detection. The final payload is a .NET infostealer that steals credentials, while Anthropic and the legitimate Claude Code installation process were not compromised.
