Security | Threat Detection | Cyberattacks | DevSecOps | Compliance

Zero Trust was built to fix an older assumption: if you were inside the network, you were trusted. Then, Cloud, SaaS and remote work broke that, so security moved toward identity, device checks, MFA, least privilege, and continuous verification. But now, with agents, the messy bit starts after access. The agent reads a prompt, pulls context, chooses a tool, calls an API, and may trigger a workflow. The login tells you the agent is “trusted”.

Cato AI Labs has discovered two critical remote code execution (RCE) vulnerabilities in Cursor IDE, the popular development environment which, according to Cursor, is used by over half of the Fortune 500. Both RCE vulnerabilities, which we refer to as “DuneSlide,” achieved a 9.8 CVSS score, and involve breaking out of the IDE’s sandbox environment and were assigned CVE IDs CVE-2026-50548 and CVE-2026-50549.

AI did not create data exposure. It changed the consequences of it. Sensitive data, excessive permissions, and broad access policies have long existed across cloud environments. In the AI era, those issues are no longer passive governance concerns. They directly influence what AI users, copilots, agents, and applications can access, process, and expose. AI has turned a posture problem into operational risk. The challenge is not simply that more data is available.

Despite three forum seizures, five administrator arrests across three operations, and the conviction of its founder, ShinyHunters remains active. The real story of ShinyHunters in 2026 is not just persistence, but the evolution of a cybercrime brand that adapts faster than defenders and law enforcement can respond. The 2025–2026 tactics make this persistence especially dangerous. Organizations using Salesforce, Salesloft Drift, Gainsight, or similar third-party SaaS integrations are at risk.

Agentic AI is evolving from assistants that answer questions into systems that can remember, use tools, call APIs, interact with SaaS applications, and improve over time. Hermes Agent, developed by Nous Research, reflects this shift as a self-improving agent that can create skills, persist knowledge, and build context across sessions., reflects this shift as a self-improving agent that can create skills, persist knowledge, and build context across sessions.

Cato CTRL recently analyzed an operator’s command-and-control (C2) server’s entire 33 days operation, including the steps he took to preserve access after the takedown. 339 commands. Four French victims. Between March 30 and May 1, 2026, Cato CTRL studied every command issued by a French-speaking threat actor (“Poisson”) against one French automotive small business and four French individuals.

A CVE lands in the morning. Hours later, attackers are exploiting it in the wild. The patch is not ready, the change window is days away, and the clock is already running. None of this is new. What changed is that vulnerability exploitation is now the most common path into organizations.

Modern IT and security teams no longer evaluate platforms in isolation. They ask how a platform fits into the architecture they run, the workflows they trust, and the outcomes they need to improve. Enterprise stacks are not isolated; they are interdependent. Identity shapes access, endpoint posture influences policy, while SIEM tools drive investigations and rely on shared data and context. AI tools introduce new layers and patterns of usage, risk, and data movement across the network.

Cato CTRL researchers recently identified an undocumented, active phishing campaign targeting Brazilian organizations with fake business-document lures, downloading a NinjaOne Remote Monitoring and Management (RMM) agent. The use of NinjaOne is particularly significant, underscoring how attackers no longer need exotic malware to penetrate an enterprise. Familiar business workflows and software is enough.