In a groundbreaking development, a team of Carnegie Mellon University researchers has demonstrated that large language models (LLMs) are capable of autonomously planning and executing complex network attacks, shedding light on emerging capabilities of foundation models and their implications for cybersecurity research.
The project, led by Ph.D. candidate Brian SingerOpens in new window, a Ph.D. candidate in electrical and computer engineering (ECE)Opens in new window, explores how LLMs—when equipped with structured abstractions and integrated into a hierarchical system of agents—can function not merely as passive tools, but as active, autonomous red team agents capable of coordinating and executing multi-step cyberattacks without detailed human instruction.
Read more…
Source: Carnegie Mellon University
Sign up for the Cyber Security Review Newsletter
The latest cyber security news and insights delivered right to your inbox
Related:
- Copy Fail vulnerability allows attackers to gain root access on virtually any modern Linux distribution
April 30, 2026
A working exploit written in Python (later released in other programming languages as well) consists of about ten lines of code and uses standard system calls that are indistinguishable from normal system activity. Kaspersky explain what the CVE-2026-31431 vulnerability, unofficially named as Copy Fail and published on April 29, is. Kaspersky also have some advice on ...
- Inside Shadow-Earth-053: A China-Aligned Cyberespionage Campaign Against Government and Defense Sectors in Asia
April 30, 2026
Through ongoing analysis of ShadowPad implants targeting South and Southeast Asia, TrendAI Research has uncovered a series of new related campaigns that are tracked under a temporary intrusion set (a provisional cluster of related activity pending formal attribution) designated SHADOW-EARTH-053, which we assess to be aligned with China’s broader strategic interests. Trend Micro telemetry indicates that ...
- Hackers stole hundreds of thousands of Roblox accounts
April 30, 2026
More than 610,000 Roblox accounts were reportedly stolen. Was yours or your child’s among them? Ukrainian police arrested three individuals in Lviv who allegedly orchestrated one of the largest Roblox account theft operations to date. Between October 2025 and January 2026, the hacking group is said to have compromised over 610,000 Roblox accounts, including at least 357 ...
- Nasty cPanel vulnerability probably exploited as a 0-day
April 30, 2026
Emergency patches are available for a critical vulnerability in cPanel and WHM that allows attackers to bypass authentication and gain root access to servers managed using it. Given that cPanel and WebHost Manager (WHM) control panel help manage properties for 70 million domains, by some estimates, and the critical severity of CVE-2026-41940 (9.8), the vulnerability is ...
- CISA flags data-theft bug in NSA-built OT networking tool
April 29, 2026
The Cybersecurity and Infrastructure Security Agency (CISA) is warning anyone who uses GrassMarlin, a tool developed by the National Security Agency (NSA), about a new vulnerability that attackers can use to snoop on sensitive information. First reported by Grady DeRosa, senior industrial pentester at Dragos, the weak spot affects all versions of GrassMarlin, a tool developed ...
- Have I Been Pwned claims Pitney Bowes hit by 8.2M email address leak
April 29, 2026
Logistics technology company Pitney Bowes, which makes franking machines for US postage, is the latest scalp claimed by ShinyHunters and its ongoing spree of pay-or-leak attacks against major organizations. Data breach tracker Have I Been Pwned (HIBP) confirmed the breach on April 27, with 8.2 million unique email addresses included in the dump alongside names, phone ...