UC Santa Cruz

In this section we highlight various fields in JavaScript that an attacker can use to transmit information. We distinguish between immediate channels where modifying this fields sends information to an attacker and delayed channels were no information is sent to the attacker until the user performs some action.

• href attribute of links (<a>)
• form submission (elements of the form)

• cross-site loading tags (in the DOM)
  • script (src attribute)
  • link (href attribute), for stylesheets
  • img (src attribute)
  • video (src attribute or source elements)
  • audio (src attribute or source elements)
  • object/embed (flash, java applets, etc.)
  • ifame (src attribute)
• XmlHttpRequest (XHR)
• form.submit
• postMessage

For storage channels data is not directly transmitted to an outside source. However, any security labels need to be persisted along with the data or these need to be treated like delayed channels.

• cookies
• window.name
• sessionStorage (html5 web storage)
• localStorage (html5 web storage)
• persistence mechanisms for flash (or other plugins)

For confidentiality, we assume that each origin may be its own security principal. We note that 'http://ucsc.edu' and 'http://soe.ucsc.edu' are different principals, as are 'http://ucsc.edu' and 'https://ucsc.edu'. In addition we have a LOCAL_ONLY principal for data that should never leave the client.

In order to safely handle exfiltration attacks, we will also need a notion of integrity. For all data, we will need to track which principals have influenced the data. (Note that without declassification, the moment that more than one principal have affected any confidential data item, it becomes LOCAL_ONLY).

• passwords – allow sending to submission origin but only if not influenced by another principal. Perhaps also restrict output channels to form submission and XHR?
• geolocation – LOCAL_ONLY, unless the user authorizes its release (this would allow us to possibly securely loosen the current restrictions on geolocation).
• HTML elements/attributes marked as confidential by the website's developers.

Since XHR can be used to pull sensitive data from different pages we need to handle its results with some care.

• By default SSL secured resources are treated as confidential to the origin.
• By default other results from XHR are treated as public
• Both of the above cases may be changed by the developer via HTTP headers

In this section we discuss a possible approach to limit a scripts access to the DOM.

• an external configuration file specifies which external scripts have access to which DOM elements
• local script files and inline JavaScript have no restriction to the DOM
  • note that this will not stop all forms of XSS attacks but might limit the number of characters available to write the attacking script (e.g. the DB fields with size limits)

• analytics – do not need access to the DOM?
• most advertisements – need only modify/select specific DOM elements
• embedded maps, videos, etc. – generally only modify specific DIV/SPAN elements
• target-word adds – need extensive access to the DOM (example can be seen at: http://www.tomshardware.com/reviews/core-i3-530-overclock-lga-1156,2626-4.html)

Exfiltration attacks present a significant threat to confidentiality for information flow analysis. In short, exfiltration attacks involve sending confidential information back to its origin, but to another account. (For example, an attacker might send someone else's confidential information from Facebook back to his own Facebook account).

In order to defend against these attacks, we will need to add a notion of integrity. Confidential data can only be sent back to its origin if the decision to do so was not influenced by any other principal. (This solution is one proposed for safe declassification).

We note that cross-site loading tags (e.g. image tags) might not require this restriction, since it does not seem likely that a site would use these tags to pass information between accounts.