AT&T sent a explanatory letter to its iPad subscribers Sunday, even as new details emerged that last week's leak of customer information was not as difficult as the company maintains, and, contrary to the company's claims, could be used to gain additional data.
The June 13 letter to "Dear Valued AT&T Customer" carried the signature of Dorothy Attwood, senior vice president, public policy, and chief privacy officer for AT&T.
"On June 7 we learned that unauthorized computer 'hackers' maliciously exploited a function designed to make your iPad log-in process faster…," according to the Attwood e-mail. "Within hours, AT&T disabled the mechanism that automatically populated the email address. Now, the authentication page log-in screen requires the user to enter both their email address and their password."
According to the company's e-mail, "I want to assure you that the email address and ICC-ID were the only information that was accessible. Your password, account information, the contents of your email, and any other personal information were never at risk. The hackers never had access to AT&T communications or data networks, or your iPad."
But both the original hackers, a group called Goatse Security, and two other security specialists say that AT&T's claim may not be entirely accurate.
The original exploit crafted by Goatse simply took advantage of a feature in AT&T's Web-based log-in process. Goatse noticed that when an iPad user started to log in, the integrated circuit card identification (ICC-ID) number, which is associated with the iPad's SIM card, caused the Web site to return the log-in screen with that user's e-mail address. The hackers wrote some code and a script to use the guess-able ICC-ID number sequences, and then repeatedly queried the AT&T Website to harvest over 100,000 e-mail addresses of iPad 3G users.
But the ICC-ID can be used to determine another bit of information, the International Mobile Subscriber Identity (IMSI) of each affected user, according to two security experts cited in a story by eWeek.
An attacker can use an "IMSI catcher" to scan an area for mobile cellular devices. When one is found, the known IMSI could then be associated with it, helping to find the location of a known user, according to the report.
From the eWeek story: "You can do this without knowing the IMSIs of people, but you won't know which IMSI belongs to which user," explained independent security researcher Nick DePetrillo. "There are other ways to determine that, but knowing ahead of time also helps, like in the case of the AT&T leak."
The iPad information could be used in mounting a much more technically difficult assault on information that's held on the "non-3G data portions of the GSM network," according to both DePetrillo and Don Bailey, a consultant with iSec Partners, a security consultancy in San Francisco.
"Through IMSI catching, an attacker could potentially intercept control messages or other data that might not be protected by the stronger encryption of the 3G data network.... [But] there is no known way to directly compromise or take control of a user's iPad with this information."
An AT&T spokesman declined comment on these new claims.
The AT&T letter triggered a harsh response from Goatse Security, in the form of a blog post by one its members, Escher Auernheimer.
"If not for our firm talking about the exploit to third parties who subsequently notified them, they would have never fixed it and it would likely be exploited by the RBN [ Russian Business Network, a cybercrime organization] or the Chinese, or some other criminal organization or government (if it wasn't already)," Auernheimer wrote.
Such potential still exists, he insisted, because Apple has failed to fix a basic weakness in the iPad's Safari browser, and both Apple and AT&T have failed to notify users of this vulnerability.
The exploit is a semantic integer overflow, which Goatse documented in a March 2010 posting. "This bug we crafted allows the viewer of a webpage to become a proxy (behind corporate and government firewalls!) for spamming, exploit payloads, password brute-force attacks and other undesirables," Auernheimer writes. "The kicker is that this attack cannot be detected by any current IDS/IPS system." According to Auernheimer, Apple patched the desktop Safari browser but still has not patched the iPad version.
Auernheimer has not responded to an e-mail with follow-up questions.