|
|
|
|
| |
Credit:
The information has been provided by Inferno.
The original article can be found at: http://securethoughts.com/2009/06/phishing-with-url-obfuscation-continues-in -safari-4
|
| |
Vulnerable Systems:
* Safari 4 version 4.0.2 and prior
Immune Systems:
* Safari 4 version 4.0.3
It is possible for a malicious website to place arbitrary sites into your Top Sites view through automated actions. The attack technique makes use of javascript windows where in a small window is used to repeatedly browse to different sites that the attacker wants to add in your Top Sites list. This window is completely hidden using the window.blur function and user won't know that is happening in the background. Please note that this attack is not possible using invisible iframes as Safari does not use iframe urls to decide Top Sites content.
Once the attack completes execution, the small window gets closed and the next time you use Safari Top Sites, it will be have the attacker's defined sites replace your existing legitimate sites. To make this decision of which sites to replace with, an attacker can first use the CSS History Hack found by Jeremiah Grossman[2] and then accordingly set fake sites relative to those user's visited websites. Hence, this could easily facilitate a serious phishing attack. The situation is worsened by the Safari's inadequate protection against URL obfuscation attacks as highlighted in [3], which makes it almost impossible for a regular user to spot the fake site and differentiate it from a legitimate one.
The two input parameters in this attack are the number of times the fake website should be visited (n)(default=28) and timeout(t)(default=2 sec) that triggers a switch between two fake websites. It is very simple and adds two fake websites for bankofamerica.com and gmail.com to your top sites. (it does not check your browser history, but that is left as an exercise for the reader :)). Also, you might have to increase the parameter value of 'n' if you visit your favorite sites very often.
A real-world hacking scenario would look like:
1. Attacker injects malicious javascript on
(a) His or her evil site OR
(b) On a legitimate site which allows javascript (e.g. bulletin boards, dashboards, etc).
2. Victim visits the above site.
3. Malicious javascript runs and first checks browser history (using CSS history hack[2]) from a list of Alexa Top 500.
4. Attacker replaces the user's visited sites with fake phishing sites (makes legitimate sounding names with url obfuscation).
5. Every time user opens a phishing site and gets a login page, user's credentials gets stolen. Attacker will present a login error message, asking user to try again later. At the same time, attacker will reset that phishing site back to the legitimate page. This way, user will never know what happened.
6. On another note, attacker can always keep atleast 1 or 2 phishing websites at all times in Top Sites. This will help the attacker to maintain persistent control of a user's session and every time user visits a new site, it will be detected by the attacker and will be replaced by a phishing site in Top Sites.
CVE Information:
CVE-2009-2196
Disclosure Timeline:
May 21, 2009: Vulnerability discovered by Inferno.
May 21, 2009: Apple contacted.
May 21, 2009: Automated response from Apple.
May 26, 2009: First response from Apple Security Team.
Jun 03, 2009: First Status update provided by Apple.
Jun 27, 2009: Second Status update provided by Apple.
Jul 24, 2009: Coordinated public release of Advisory with Apple.
Aug 11, 2009: Software Update and Public Advisory issued by Apple.
|
|
|
|
|
