Browsing: AI agent security risks and attack vectors