Promptware Elevates LLM Attack Campaign Severity

On Feb. 3, 2026, security researchers Ben Nassi, Bruce Schneier, and Oleg Brodt warn that “promptware” — multi-stage attacks against LLM-based workplace tools — is increasing. They describe deliberate campaigns that use deceptive prompts to gain entry, escalate privileges, establish persistence, and move laterally across services, and they urge organizations to apply multi-phase cybersecurity controls to interrupt these attacks.