Autonomous attack agent in ‘Skynet mode’ hijacks other models to carry out its evil deeds
Threat actors built an AI agent designed to hijack exposed AI models to run autonomous attacks, with instructions allowing it to carry out up to
Threat actors built an AI agent designed to hijack exposed AI models to run autonomous attacks, with instructions allowing it to carry out up to
security Bad things (and software bugs?) come in threes As the PHP2Shell chaos continues, researchers have found a third PHP vulnerability that could allow attackers
ai and ml Monitoring was only partial and was easily bypassed An independent evaluation of OpenAI’s internal misalignment monitoring system revealed that the system was
Linus Torvalds is embracing AI to write Linux kernel code, and some people are having fits. He has a message for them. So do I.
networks Whitehall wants tougher penalties and new powers over telecoms operators as Russian activity raises alarm Britain is threatening subsea cable saboteurs with up to
security “Adversaries are clearly spending time (and tokens) on this bug,” researcher says. Researchers have developed a new exploit that shows how even Address Space Layout
security Vendor ‘strongly recommends’ upgrading to a fixed version ‘as soon as possible’ GitLab on Wednesday patched a maximum-severity flaw that could allow unauthenticated attackers
security We didn’t want to issue SSL/TLS certs anyway, Indian firm suggests after one private key turns out to be the number 4 After embarrassing
security Ministers report 32 unavailable and four partly disrupted, but no confirmed compromise The PHP2Shell security fiasco has forced 32 UK government services offline and
In a world full of cyber risk, account takeover has emerged as one of the most insidious of all threats. If an attacker can successfully
cyber-crime Cloudflare reports millions of exploit attempts, while Amazon spotted attackers trying their luck four days before patches landed Attackers are hammering servers with millions
security AI agents did the dirty work, with hackers gaining privileged access in some environments in under three hours China-aligned attackers used AI agents and
security PWNED Misaligned incentives led to an intractable argument Welcome back to PWNED, the column where we talk about companies and organizations that leave the
Executive summary This report analyzes activity during which UAC‑0099 gained initial access on July 16th, 2025 and handed it off to Sandworm, which then conducted
security Says it passed audits and made just a minor mystery error Indian certification authority eMudhra has defended the laughably weak private keys it used
PHP developers have accused software supply chain security shop Chainguard of effectively revealing critical bugs in the open-source scripting language ahead of a planned official
Unknown crims compromised LiteLLM’s release process on Wednesday, distributing a malicious version (1.107.0) containing malware designed to steal cloud and AI credentials and a remote
security Careful what you say after goodbye, or better yet, update your messaging app Make sure you don’t say anything mean or secret after hanging