Security | Threat Detection | Cyberattacks | DevSecOps | Compliance

CrowdStrike OverWatch identified a VAPOR PANDA campaign using ARTEX alongside the Chinese-developed DeepSeek large language model (LLM) and a second commercial LLM to conduct intrusions against multiple high-value targets in parallel.

Prompt injection started as a simple idea during its initial chatbot phase: Tell the model to ignore its instructions and return information in violation of its guardrails. Now, prompt injection attacks can aim at something harder to see: how an agent plans, and how a model reasons and decides which rules apply.

CrowdStrike researchers asked a frontier AI model to explain a function in a small C web server. While it had previously explained every other function in the codebase without hesitation, it refused to explain this one. The request was routine. We'd only asked what the code does; we hadn't asked about security, exploits, or vulnerabilities. But a single token in the code had changed: a max swapped for a min; a one-character logic inversion that creates an out-of-bounds write.

Cyber intrusions targeting the legal sector nearly doubled between 2023 and 2026, reflecting the value of legal entities to both eCrime and nation-state adversaries. Legal entities hold privileged communications, corporate transactions, intellectual property, government matters, and financial records that make them valuable targets for threat actors seeking financial gain or long-term strategic intelligence.

AI is changing from a tool that generates content into agents that reason, decide, and take autonomous action. Employees now use AI assistants in their browsers and desktop apps. Developers run coding agents that read files and execute commands on their machines. Business teams build agents and connect them to enterprise data through Model Context Protocol (MCP) servers. This shift from asking AI to tasking AI creates hard questions for security teams: Which AI agents are operating in the environment?

Two critical memory-corruption remote code execution (RCE) vulnerabilities, collectively dubbed PHP2Shell, have been disclosed in core PHP, the underlying engine powering a vast portion of the web. The CrowdStrike Falcon platform provides out-of-the-box visibility and detection coverage for post-exploitation activities of both vulnerabilities.

A new report from the CrowdStrike Counter Adversary Operations team details research into four previously unreported threat actors, each of whom left their offensive agent frameworks exposed via open directories. In doing so, they shared detailed insight into their use of AI agents and revealed how adversaries are quickly becoming more proficient in AI.

Some cyber incidents leave no margin for error, whether that's an attack that interrupts energy, water, or financial services, or a breach that triggers reporting obligations in several jurisdictions within hours. The UK National Cyber Security Centre (NCSC) created the Enhanced Level of its Cyber Incident Response (CIR) scheme for incidents like these. It recognizes providers that meet rigorous standards for high-quality incident response, at the highest level of risk.

In late September, 2026, CrowdStrike's Counter Adversary Operations team collaborated with the United States Federal Bureau of Investigation (FBI) to disrupt several laptop farms operating across the U.S. maintained by DPRK-nexus adversary FAMOUS CHOLLIMA to fraudulently employ operatives at American companies while concealing their true location and identity.

Critical infrastructure is where cyber risk becomes real-world risk. The environments powering energy, water, manufacturing, and transportation are among the most complex to defend as they often combine decades-old operational technology (OT) with modern IT. Downtime is not an option, and patching isn’t always possible. Now, AI is raising the stakes. Adversaries are using AI to move faster and operate at greater scale, compressing the time defenders have to respond to attacks.