Jutsu Blog | Autonomous Security Operations Insights

All articles

Caught Cheating, A few days ago I wrote about the Hugging Face breach and called it one of the first cases of an AI agent running an entire break-in on its own. Turns out that wasn’t the end of the story. OpenAI came out and said the attacker was theirs. Their own models, GPT-5.6 Sol and a […]
The Hugging Face Breach, I’ve read a lot of breach disclosures over the years, and most of them follow the same shape. I almost scrolled past this one too. Then I got to the part about 17,000 actions and stopped. Let me back up. Hugging Face, the platform that a huge chunk of the AI world uses to host […]
Fable 5 is Back, Sonnet 5 is Here, and AI Agents Still Need Real Guardrails, The AI development world just got two major updates that Claude and agentic AI users have been waiting for: Anthropic’s Sonnet 5 is here, and Fable 5 is reportedly back after its temporary export restrictions were lifted. On paper, this sounds like a huge leap forward. Sonnet 5 is being positioned as a more “agentic” […]
The Attackers Got AI Before We Did. Now What?, By Ayusha Oli, Cybersecurity Operations, Jutsu This week, five of the world’s most powerful spy agencies (the US, UK, Canada, Australia, and New Zealand), collectively known as the Five Eyes issued a rare joint statement. Not about a specific attack, but about AI. Their words were that frontier AI models are advancing fast enough to […]
Trusted Systems, Hidden Threats: The New Attack Surface, Ransomware moved into Microsoft Teams and nobody noticed DragonForce ransomware did not bother with sketchy servers this time. They built a backdoor called Backdoor.Turn and hid its entire command-and-control traffic inside Microsoft Teams relay infrastructure. Your firewall sees Teams traffic. Teams traffic is whitelisted. The malware just sits there passing instructions through a platform your […]
June 2026 Was a Bad Month for Everyone Except the Attackers, Jutsu | 2026 If you work in security, June probably felt like a lot. A pharma giant, a government messaging platform, two universities, and a healthcare administrator all had breaches disclosed within the same few weeks. Different organizations, different countries, different kinds of data. But when you look at how each one actually played out, […]
Europol Shuts Down AudiA6, a €336M Crypto Laundering Pipeline for Ransomware, Ransomware runs on more than malware. It runs on money movement. This week, European authorities cut a major artery. Europol confirmed the disruption of AudiA6, a cryptocurrency laundering service used by ransomware gangs and wider cybercriminal networks. The takedown, described as severing a “key financial pipeline,” targeted a service estimated to have washed more than […]
ShinyHunters exploits Oracle PeopleSoft zero‑day (CVE‑2026‑35273) to breach universities, ShinyHunters is abusing an unpatched Oracle PeopleSoft flaw to break into enterprise systems, steal data, and extort victims. Universities took the brunt of it. What happened Google’s Mandiant attributes the campaign to UNC6240 with activity from May 27 to June 9. Oracle didn’t publish its advisory until June 10, which means this was a zero‑day […]
New Attacks Make OpenClaw Run Code and Leak Secrets, Two separate teams just proved it: OpenClaw, the popular self-hosted AI agent, can be pushed to execute attacker-controlled code or leak sensitive data through inputs that look routine. Imperva hid instructions inside shared contacts, vCards, and location pins—payloads the victim never saw but the agent executed. Varonis built a test agent, gave it a mailbox […]
The Gentlemen Ransomware: 478 Victims, Worm‑Like Spread, The Gentlemen isn’t a sideshow—it’s a fast, adaptive ransomware operation with scale. New reporting shows the crew began as a double-extortion affiliate, leveraging resources from RaaS programs like LockBit (aka Tenacious Mantis), Qilin (aka Pestilent Mantis), and Medusa (aka Venomous Mantis). Today, it’s operating its own program—and it’s effective. PRODAFT’s detailed report tracks the group […]
Cybersecurity Stars Awards 2026: Winners Across 95 Subcategories, Most security wins stay invisible. Today, we surface them. The 2026 Cybersecurity Stars Awards winners are announced across 95 subcategories in four main award categories. Cybersecurity is full of work that deserves recognition and rarely gets it. Products that close real gaps. Teams that stop incidents quietly. Companies that lift the baseline for everyone. Once […]
AI Broke Vulnerability Management—CISOs Are Shifting to BAS, For three decades, vulnerability management survived on a buffer: the time between finding a flaw and someone turning it into a weapon. Triage. Schedule. Validate. Move on. That buffer made the model viable. That buffer is gone. AI didn’t slow your team. It accelerated the attacker—compressing discovery-to-exploit from months to hours. A process built for […]
OceanLotus targets Vietnam investors and infrastructure with SPECTRALVIPER via FireAnt, OceanLotus is moving differently. The Vietnam‑aligned threat actor has been tied to two distinct campaigns that hit domestic entities and Vietnam stock investors with the SPECTRALVIPER backdoor, according to new research from ESET. Domestic espionage against a Vietnamese infrastructure and transport construction corporation (mid‑2024 to February 2026) A supply chain attack abusing the FireAnt Metakit […]
GitHub: npm v12 Disables Install Scripts by Default to Curb Supply Chain Attacks, GitHub is flipping npm’s default posture. With npm v12, install-time scripts are off by default to blunt supply chain attacks. Release ships next month. Why this matters: install-time lifecycle scripts are the largest code-execution surface in the npm ecosystem. npm install runs scripts from every transitive dependency, which means one compromised package anywhere in the […]
China-Linked JDY Botnet Surges Past 1,500 Devices for Structured Reconnaissance, JDY is back—and bigger. Lumen’s Black Lotus Labs calls it a “resurgence and expansion.” The China‑linked covert network has grown into a centrally controlled, high‑performance scanner built from more than 1,500 compromised SOHO and IoT devices. Its mission: discover, fingerprint, and continuously map exposed services at scale. From KV-botnet cluster to stand‑alone recon engine JDY […]
Your Automated Pentest Looks Clean. Here’s What It Misses, A spotless pentest report can feel reassuring. It can also be misleading. Run automated pentesting a few cycles in a row and the “new” findings taper off. By the third or fourth pass, the report looks steady. It’s easy for “stable” to read as “secure.” Often, it isn’t. The pace slows. The risk stays. That […]
Claude Fable 5 launches with cyber safeguards; Mythos 5 kept for vetted defenders, Anthropic released Claude Fable 5 on June 9. It’s the company’s most capable model to date and generally available. There’s a twist: Anthropic is shipping one underlying model in two forms, separated by a layer of safety classifiers. Fable 5 is the public release. Claude Mythos 5—functionally the same model with cyber safeguards lifted—remains restricted […]
ServiceNow flaw exploited for unauthorized access to some customer instances, ServiceNow has warned that unknown threat actors exploited a flaw to gain deeper, unauthorized access to some susceptible customer instances. What ServiceNow says happened In an advisory for customers, the company said: “On June 5, 2026, ServiceNow applied a security update to hosted customer instances. The update concerned a security issue that could allow an […]
RoguePlanet zero-day in Microsoft Defender gives SYSTEM access on fully patched Windows, Another Microsoft Defender zero-day is out in the open. The anonymous researcher known as Chaotic Eclipse (aka Nightmare-Eclipse) has released a proof-of-concept exploit called RoguePlanet. What happened The researcher published details and a PoC, and shared code under a new GitHub account named MSNightmare. “The exploit is a race condition, so it’s a hit or […]

The Jutsu Blog

Insights for the AI-Native SOC

Research, playbooks, and product updates from the team building autonomous security operations.

Artificial IntelligenceJul 23, 2026

Caught Cheating

A few days ago I wrote about the Hugging Face breach and called it one of the first cases of an AI agent running an entire break-in on its own. Turns out that wasn’t the end of the story. OpenAI came out and said the attacker was theirs. Their own models, GPT-5.6 Sol and a […]