I understand that during the 'Implicit Grant' flow, the Authorization server provides the Access Token to the Client (running in UserAgent/Browser) via the URL.
Wouldn't the hacker having access to the UserAgent be able to decipher/extract the 'Access Token' and use it for the resource calls ?
How do we prevent it from happening ?