🛡️ InfoSec Blue Team Briefing

Friday, November 20, 2026

🎧 Audio Briefing

Download MP3

BlueTeamSec briefing for Friday the 20th of November 2026, pulled from what the community was sharing on infosec.pub. We've got 10 stories to get through, so let's crack on.

Google has a case study on COLD RELIC, a Russian state-sponsored espionage actor that Google tracks, also known as Star Blizzard. It chains compromises, using stolen directories and credentials from one victim to impersonate trusted people and reach the next, and it registers rogue MFA devices and passkeys that survive password resets. The defensive advice is solid: phishing-resistant MFA and revoking sessions and OAuth grants, not just resetting passwords. Related Russia-linked coverage on the 18th was the EU sanctions announcement.

The US Department of Justice has indicted Igor D'iu, a Russian national accused of acting as an unregistered foreign agent. Prosecutors say he was recruited over Telegram to surveil employees of a Palo Alto technology and defence company, using a forged identity and setting up wireless equipment near the office to potentially intercept device data. It's a physical-world reminder that the threat model for staff at defence suppliers doesn't stop at the firewall.

Cisco Talos is tracking UAT-12643, a financially motivated cluster it assesses with low confidence to be North Korea-aligned. It goes after crypto holders, gaming platform users and VoIP software operators, and its Go-based remote access tool hides its command server addresses in the Solana blockchain. It also persists through the Windows recovery environment and has a kill flag that blue-screens the machine. That's a rather rude way to say goodbye.

AttackIQ's research, called Weighing Cyber Power, maps around 20 Iranian state-sponsored groups and describes a handoff model. Espionage-focused teams get into networks, then pass that access to more aggressive operators who can turn quiet intelligence gathering into disruption or destruction. Worth a look if you've been treating a spy-grade intrusion as the lesser problem.

The French regulator CNIL has fined an unnamed company five million euros after an audit found customer passwords stored in plain text, shared credentials for database access, and contact form data kept for over ten years. There's no breach or threat actor involved, and CNIL published the decision for educational purposes. Salted hashing and individual accounts in 2026 are apparently still a teaching moment.

Broadcom describes an attack on a major African stock exchange that lasted just three minutes and twelve seconds, using 41 apparently AI-generated PowerShell modules to harvest secrets and exfiltrate data through DNS lookups. Six persistence implants on the server were attributed with high confidence to Witchetty, which is a China-linked espionage group according to the report, though where the AI-written toolkit came from is unconfirmed. Another entry in the AI-changes-everything file, though this one at least comes with evidence.

Dutch police report that unauthorised third parties stole identifying information on hundreds of European Revolut customers in September, and criminals then used it to plan violent attacks in Dutch cities including The Hague and Amsterdam. Police foiled the attempts and arrested ten suspects, while the investigation into how the breach happened continues. It's a grim illustration of what happens when stolen customer data ends up in the wrong hands.

Censys has investigated a supply chain poisoning campaign where malicious JavaScript was injected into third-party libraries from Optimizely, Kissmetrics, StorifyMe and minute-ly. Sites loading them, including the BBC, Zoom, CNN and GitLab, served a multi-stage payload that steals data and attempts iOS exploitation. Censys found the control panel, the Caishen iOS exploit backend and crypto storefronts across 142, 144 and 55 web properties respectively. Two entries covered the same research, and it follows Hudson Rock's write-up on the 18th of poisoned scripts on those same big sites. Your third-party script inventory is probably longer than you think.

Netskope reports that the actor ARTEX has built an automated AI-driven exploit skill called PHP2Shell, which turns file uploads into remote code execution through PHP's default-enabled upload progress feature. It's a critical-rated flaw, and roughly 27 million PHP installations are exposed, 22.8 million of them on end-of-life versions. We covered ARTEX on the 16th and a detection pack for its misuse on the 19th, so this is a continuing story. If you have old PHP lying around, the question is what's been uploaded already.

That's the lot for Friday from BlueTeamSec. Where we linked activity to a country or group, that was the original authors' assessment, not ours, and the analysis was automated. Have a good weekend.

📰 Articles Covered