Ciberseguridad, vulnerabilidades y privacidad
A three-person security team used frontier AI to turn a malformed image upload into a route toward OpenAI’s internal systems in less than 72 hours—but the humans still drove every consequential move.
Unbound versions through 1.26.0 contain a critical DNSSEC validator heap overflow. NLnet Labs released version 1.26.1 with fixes for this and eight additional vulnerabilities.
Cisco says attackers are exploiting a Secure Email Gateway flaw that can let unauthenticated attackers execute commands as root.
Researchers report that Mantax Otax can steal OTPs, monitor activity, encrypt files and harass victims through sideloaded Android APKs.
Anthropic says Russia-linked developers used Claude Code to prototype swarm coordination, targeting and detonation software for combat drones.
Researchers say the DDRop hardware attack can silently drop memory writes on protected cloud servers, reusing older encrypted data without triggering the encryption checks.
Microsoft has reported two significant phishing campaigns targeting cloud accounts, utilizing advanced social engineering techniques and passkey-themed tactics.
Three vulnerabilities in JFrog Artifactory are currently being exploited, with patches available for all. Organizations are urged to upgrade to secure versions promptly.
GitLab has addressed a critical vulnerability, CVE-2026-85706, that allows unauthorized file access, with reports of active exploitation occurring shortly after its disclosure.
Microsoft's recent Patch Tuesday has set a new record with 974 CVEs addressed, including two actively exploited vulnerabilities. Adobe also contributed with critical updates.
