SecurityAI OpenAI releases GPT-6 Cyber and rates its hacking skill as critical OpenAI rates its new security model GPT-6 Cyber as critical for cybersecurity, so only vetted defenders in its Daybreak Red tier get access.
SecurityAIOpen Source OpenAI says its AI agents built working root exploits for 948 Linux kernel bugs An unreleased OpenAI cyber model found and exploited 948 Linux kernel flaws in one run costing about 1.2 million dollars; patches are going to maintainers.
Security Ivanti patches two more Connect Secure zero days used in a root takeover chain Ivanti fixed two more Connect Secure flaws that attackers chained with an older bug to take root on patched VPN gateways. CISA set short deadlines.
SecurityOpen Source US cyber agency confirms attacks on PHP2Shell flaws and sets a patch deadline CISA added both PHP2Shell flaws to its list of exploited vulnerabilities. Attacks began days before the patches, so unpatched PHP servers are at risk now.
AISecurity Google launches Gemini 4 Carbon for security work but only cyber defenders get it Google's Gemini 4 Carbon finds, tests and patches software flaws. Vetted defenders get it first, because the same skills can serve attackers.
Security Hacktivists leak details of more than 41000 US Homeland Security staff A leaked dataset lists names, contacts and roles of over 41,000 DHS staff and contractors. The hackers say they got in through a single sign-on service.
Security Binance loses about six billion dollars in the largest theft on record Attackers took about 6.4 billion dollars from Binance on 29 October. Analysts and the FBI tie it to North Korea, and about half has been frozen.
SecurityOpen Source A critical GitLab flaw patched in October is now being exploited in the wild CISA added the CVSS 10 GitLab flaw CVE-2026-97229 to its exploited list on 29 October. Self-managed servers on 19.2 to 19.4 need the October patch now.
SecurityOpen Source PHP patches two critical remote code execution flaws in all supported lines PHP 8.5.12, 8.4.27, 8.3.36 and 8.2.35 fix two critical flaws that can allow unauthenticated remote code execution on servers that handle web requests.
