WebAppSec Clear Site Data
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.
demo Change the demo to new header syntax. (#28) Jun 20, 2017
performance_demo Migrated the performance demo to appengine. (#47) Jan 4, 2018
.pr-preview.json Use pr-review tool (#25) May 5, 2017
.travis.yml Add a setuptools upgrade step to .travis.yml Oct 20, 2016
CONTRIBUTING.md Linked testing policy and fixed a few links (#42) Nov 8, 2017
README.md Explainer. Feb 2, 2018
index.html Add the wildcard pseudotype. (#43) Nov 30, 2017


Explainer: Clearing Site Data

The Problem

Developers wish to have control over the data associated with their origins which users' browsers store locally on their behalf. In particular, developers wish to take reasonable steps to protect users from local attackers by ensuring that certain kinds of data are removed from a user's local machine when it is no longer necessary (for example, upon signing out of an application, or upon account deletion).

Developers have direct access to a number of storage mechanisms, which makes it possible to perform this kind of cleanup from JavaScript. window.localStorage, for example, is easily dealt with a simple call to window.localStorage.clear(). Other storage mechanisms are trickier. Cookies, for instance, exist cross-origin, might not be accessible to JavaScript, and could have path restrictions that can make them quite difficult to enumerate. The browser's various caches are even more difficult to poke at, as they're intentionally opaque.

It would be nice if browsers offered developers a mechanism that would give them power to remove the data their applications are responsible for maintaining. Ideally, developers should be able to reliably ensure the following:

  1. Data stored in an origin's client-side storage mechanisms like IndexedDB, WebSQL, Filesystem, self.localStorage, self.sessionStorage, self.caches, etc. is cleared.
  2. Cookies for an origin's host are removed.
  3. Web Workers (dedicated and shared) running for an origin are terminated.
  4. Service Workers registered for an origin are terminated and deregistered.
  5. Resources from an origin are removed from the user agent's local cache.
  6. None of the above can be bypassed by a maliciously active document that retains interesting data in memory, and rewrites it if it's cleared.

The Proposal

One way to give developers the capabilities alluded to above would be to accept a server-sent assertion that an origin's data be cleared. This could be an HTTP response header whose value specified a subset of locally stored data to be cleared: of:

Clear-Site-Data: "*"

With this kind of assertion, developers could handle a number of use cases:

  1. When a user deletes their account from a given social media application, the "Oh noes, you're gone!" page could be served along with Clear-Site-Data: "*" to clear out any and all sensitive information that may have been persisted to the user's disk.

  2. If a user signs out of a given application, it might do a more targeted cleanup, removing, for example, photos from the disk cache, while retaining cookies with interesting user preferences by serving Clear-Site-Data: "cache".

  3. In the case of catastrophic failure (perhaps an applications developers learn that their servers were compromised for some period of time), developers can reduce the risk of a persistent client-side XSS by clearing out local sources of data: Clear-Site-Data: "*".