Huntress just uncovered a scary, uncommon intrusion path: the attackers don’t drop malware straight onto Windows. Instead, they aim at an Oracle database first, use SQL injection to take control, then weaponize Oracle itself to pivot into SYSTEM-level access on the target host. How it starts: an external-facing enterprise app API suffers from weak input validation. That leaves a classic SQL injection opening, letting attackers gain control of the DB. The twist: after the injection, they don’t only read or modify data. They plant a post-exploitation toolkit (khunt) inside Oracle using database objects—then trigger execution through Oracle’s built-in Java virtual machine, OJVM. At the core is OJVM: attackers use SQL statements like CREATE JAVA SOURCE to write and compile Java code directly into the database engine. Once compiled, later SQL calls can execute actions mapped to multiple khunt components. In real attacks, khuntCmd helps launch Windows cmd.exe for RCE, escalate privileges to SYSTEM, then exfiltrate sensitive Registry hives (SAM/SECURITY) for password hash recovery. The big defense lesson: stop SQLi with parameterized queries + input sanitization, and lock down Oracle accounts to limit post-injection write/execute power. #OracleSecurity #SQLInjection #OJVM #RansomwareDefense #EDRBypass #ThreatIntel
Want to learn more? Visit Explore the world, stay updated on travel insights and international affairs, and discover authentic stories from real life
评论
发表评论