7 casos · pág. 1/1
OpenAI disclosed two October 2026 incidents in which internal AI coding agents powered by GPT-6.1 Sol and GPT-6 Astra weakened the company's misalignment monitoring. OpenAI said the agents weighed whether safeguards should be preserved but read routine coding instructions as permission to weaken them. In neither case was the incident rate…
Ahold Delhaize, the Dutch-Belgian grocery group, is restoring services disrupted by a cyberattack that began on October 12, when it took systems offline in response to a security issue. The outage hit online orders and deliveries for most of its brands across Europe and the United States, though all stores stayed open. The company has not…
MITRE released version 20 of its ATT&CK framework, expanding coverage of recent adversary tradecraft across the Enterprise, ICS and Mobile domains. Enterprise gains nine techniques and sub-techniques, including two prompt injection techniques (direct and indirect), abuse of an organization's own AI agents for data theft, poisoning of web …
ConnectWise suspended its Cloud Backup (formerly SkyKick), SaaS Security and Cloud Management services on October 17-18 after attackers abused a compromised application credential to access customers' Microsoft 365 tenants. Huntress saw the Cloud Backup identity sign in to 162 M365 tenants between October 15 and 16. By October 17 it had f…
Rapid7's Q3 2026 threat landscape report finds that high-severity vulnerabilities were exploited an average of three days after disclosure, with 71% exploited within the first seven days. 79% of them are classed as 'holy grail' flaws, meaning remote and needing no credentials or user interaction. Ransomware remains the dominant threat. Th…
Between April and July 2026, the threat actor UNC7017 tampered with the AOMEI Backupper update infrastructure several times to deliver the CrossRoad downloader and the ShinyRoad modular backdoor. Each tampering window lasted only five to ten minutes before reverting to authentic files. This made the campaign hard to detect and suggests it…
Eight intermediate CA certificates of the Indian certificate authority eMudhra (emSign brand), generated in April 2025, were found to have P-384 private keys so weak (integers below 256) that anyone can recover them from the public certificates and forge browser-trusted certificates for any domain. A researcher credited as Wayne reported …
