Detects hidden prompt injection attacks embedded in web pages — protecting your AI agent from being hijacked.