As part of our class project, we are studying the attacks that could be done using CSS Injection. In our threat model, attacker can manipulate any CSS file on the server. If attacker replaces original CSS file with a malicious CSS file and if the web-server sends this malicious CSS file to the client, can the attacker execute some code in the client?
For example, a CSS file contains information like this
#mw-indicator-mw-helplink a {
background-image: url('images/help.png');
background-image: linear-gradient(transparent, transparent), /* @embed */ url('images/help.svg');
background-repeat: no-repeat;
background-position: left center;
padding-left: 28px;
display: inline-block;
height: 24px;
line-height: 24px;
}
Can an attacker change the url to some malicious url, and execute some code in client?