New “Prompt Poaching” Attack Steals Users’ AI Conversations via Malicious Browser Extensions

**Prompt poaching**: security researchers uncovered malicious browser extensions that silently monitor AI assistant tabs, scrape or intercept user prompts and AI responses, and exfiltrate sensitive data to attacker-controlled servers; attackers have distributed cloned extensions and pushed updates to compromised legitimate extensions, exposing users and organizations to intellectual property loss, targeted phishing, and identity theft, and defenders are advised to enforce centralized extension controls, prefer official clients, and monitor outbound traffic.