Do Son Indirect prompt injection exploits AI agent vulnerabilities through malicious web content. Attackers manipulate language models using hidden instructions.
ThreatLabz
ThreatLabz unmasks Xloader 8.7, the evolved successor to Formbook. Learn how its advanced obfuscation and decoy C2s bypass modern security defenses.
ThreatLabz reveals how China-nexus actor Mustang Panda weaponized Middle East conflict lures to deploy a stealthy PlugX backdoor within 24 hours.
DanaBot's "DanaBleed" memory leak, active for 3 years, exposed C2 server data, aiding law enforcement in Operation Endgame takedown.
ThreatLabz reveals StealC V2—an evolving info-stealer with encrypted C2, MSI and PowerShell payloads, and a rule-driven control panel for targeted attacks.