When Agents Attack: How Prelude's Engineering Team Fights Back
Originally published on the Prelude blog on August 24, 2026, written by Rowan Haddad. Reposted here in full. Summary When Anthropic disclosed in July 2026 that three of its own Claude models had breached real organizations during security evaluations, it confirmed something Prelude’s engineering team had already been watching: AI agents have changed the fraud problem, not because they are smarter than humans, but because they are faster. This post covers how Prelude detects agent-driven attacks, how the SMS pumping threat is evolving, and why Prelude is now fighting agents with agents. ...