By Kim Zetter
Aug, 03, 2006
LAS VEGAS -- A German computer security consultant has shown that he can
clone the electronic passports that the United States and other countries
are beginning to distribute this year.
The controversial e-passports contain radio frequency ID, or RFID, chips
that the U.S. State Department and others say will help thwart document
forgery. But Lukas Grunwald, a security consultant with DN-Systems in
Germany and an RFID expert, says the data in the chips is easy to copy.
"The whole passport design is totally brain damaged," Grunwald says.
"From my point of view all of these RFID passports are a huge waste of
money. They're not increasing security at all."
Grunwald plans to demonstrate the cloning technique Thursday at the Black
Hat security conference in Las Vegas.
The United States has led the charge for global e-passports because
authorities say the chip, which is digitally signed by the issuing
country, will help them distinguish between official documents and forged
ones. The United States plans to begin issuing e-passports to U.S.
citizens beginning in October. Germany has already started issuing the
Although countries have talked about encrypting data that's stored on
passport chips, this would require that a complicated infrastructure be
built first, so currently the data is not encrypted.
"And of course if you can read the data, you can clone the data and put it
in a new tag," Grunwald says.
The cloning news is confirmation for many e-passport critics that RFID
chips won't make the documents more secure.
"Either this guy is incredible or this technology is unbelievably stupid,"
says Gus Hosein, a visiting fellow in information systems at the London
School of Economics and Political Science and senior fellow at Privacy
International, a U.K.-based group that opposes the use of RFID chips in
"I think it's a combination of the two," Hosein says. "Is this what the
best and the brightest of the world could come up with? Or is this what
happens when you do policy laundering and you get a bunch of bureaucrats
making decisions about technologies they don't understand?"
Grunwald says it took him only two weeks to figure out how to clone the
passport chip. Most of that time he spent reading the standards for
e-passports that are posted on a website for the International Civil
Aviation Organization, a United Nations body that developed the standard.
He tested the attack on a new European Union German passport, but the
method would work on any country's e-passport, since all of them will be
adhering to the same ICAO standard.
In a demonstration for Wired News, Grunwald placed his passport on top of
an official passport-inspection RFID reader used for border control. He
obtained the reader by ordering it from the maker -- Walluf, Germany-based
ACG Identification Technologies -- but says someone could easily make
their own for about $200 just by adding an antenna to a standard RFID
He then launched a program that border patrol stations use to read the
passports -- called Golden Reader Tool and made by secunet Security
Networks -- and within four seconds, the data from the passport chip
appeared on screen in the Golden Reader template.
Grunwald then prepared a sample blank passport page embedded with an RFID
tag by placing it on the reader -- which can also act as a writer -- and
burning in the ICAO layout, so that the basic structure of the chip
matched that of an official passport.
As the final step, he used a program that he and a partner designed two
years ago, called RFDump, to program the new chip with the copied
The result was a blank document that looks, to electronic passport
readers, like the original passport.
Although he can clone the tag, Grunwald says it's not possible, as far as
he can tell, to change data on the chip, such as the name or birth date,
without being detected. That's because the passport uses cryptographic
hashes to authenticate the data.
When he was done, he went on to clone the same passport data onto an
ordinary smartcard -- such as the kind used by corporations for access
keys -- after formatting the card's chip to the ICAO standard. He then
showed how he could trick a reader into reading the cloned chip instead of
a passport chip by placing the smartcard inside the passport between the
reader and the passport chip. Because the reader is designed to read only
one chip at a time, it read the chip nearest to it -- in the smartcard --
rather than the one embedded in the passport.
The demonstration means a terrorist whose name is on a watch list could
carry a passport with his real name and photo printed on the pages, but
with an RFID chip that contains different information cloned from someone
else's passport. Any border-screening computers that rely on the
electronic information -- instead of what's printed on the passport --
would wind up checking the wrong name.
Grunwald acknowledges, however, that such a plot could be easily thwarted
by a screener who physically examines the passport to make sure the name
and picture printed on it match the data read from the chip.
Machine-readable OCR text printed at the bottom of the passport would also
fail to match the RFID data.
Frank Moss, deputy assistant secretary of state for passport services at
the State Department, says that designers of the e-passport have long
known that the chips can be cloned and that other security safeguards in
the passport design -- such as a digital photograph of the passport holder
embedded in the data page -- would still prevent someone from using a
forged or modified passport to gain entry into the United States and other
"What this person has done is neither unexpected nor really all that
remarkable," Moss says. "(T)he chip is not in and of itself a silver
bullet.... It's an additional means of verifying that the person who is
carrying the passport is the person to whom that passport was issued by
the relevant government."
Moss also said that the United States has no plans to use fully automated
inspection systems; therefore, a physical inspection of the passport
against the data stored on the RFID chip would catch any discrepancies
between the two.
There are other countries, however, that are considering taking human
inspectors out of the loop. Australia, for one, has talked about using
automated passport inspection for selected groups of travelers, Moss says.
In addition to the danger of counterfeiting, Grunwald says that the
ability to tamper with e-passports opens up the possibility that someone
could write corrupt data to the passport RFID tag that would crash an
unprepared inspection system, or even introduce malicious code into the
backend border-screening computers. This would work, however, only if the
backend system suffers from the kind of built-in software vulnerabilities
that have made other systems so receptive to viruses and Trojan-horse
"I want to say to people that if you're using RFID passports, then please
make it secure," Grunwald says. "This is in your own interest and it's
also in my interest. If you think about cyberterrorists and nasty,
black-hat type of guys, it's a high risk.... From my point of view, it
should not be possible to clone the passport at all."
Hosein agrees. "Is this going to be the massive flaw that makes the whole
house of cards fall apart? Probably not. But I'm not entirely sure how
confident we should feel about these new passports."
Grunwald's technique requires a counterfeiter to have physical possession
of the original passport for a time. A forger could not surreptitiously
clone a passport in a traveler's pocket or purse because of a built-in
privacy feature called Basic Access Control that requires officials to
unlock a passport's RFID chip before reading it. The chip can only be
unlocked with a unique key derived from the machine-readable data printed
on the passport's page.
To produce a clone, Grunwald has to program his copycat chip to answer to
the key printed on the new passport. Alternatively, he can program the
clone to dispense with Basic Access Control, which is an optional feature
in the specification.
Grunwald's isn't the only research on e-passport problems circulating at
Black Hat. Kevin Mahaffey and John Hering of Flexilis released a video
Wednesday demonstrating that a privacy feature slated for the new
passports may not work as designed.
As planned, U.S. e-passports will contain a web of metal fiber embedded in
the front cover of the documents to shield them from unauthorized readers.
Though Basic Access Control would keep the chip from yielding useful
information to attackers, it would still announce its presence to anyone
with the right equipment. The government added the shielding after privacy
activists expressed worries that a terrorist could simply point a reader
at a crowd and identify foreign travelers.
In theory, with metal fibers in the front cover, nobody can sniff out the
presence of an e-passport that's closed. But Mahaffey and Hering
demonstrated in their video how even if a passport opens only half an inch
-- such as it might if placed in a purse or backpack -- it can reveal
itself to a reader at least two feet away.
Using a mockup e-passport modeled on the U.S. design, they showed how an
attacker could connect a hidden, improvised bomb to a reader such that it
triggers an explosion when a passport-holder comes within range.
In addition to cloning passport chips, Grunwald has been able to clone
RFID ticket cards used by students at universities to buy cafeteria meals
and add money to the balance on the cards.
He and his partners were also able to crash RFID-enabled alarm systems
designed to sound when an intruder breaks a window or door to gain entry.
Such systems require workers to pass an RFID card over a reader to turn
the system on and off. Grunwald found that by manipulating data on the
RFID chip he could crash the system, opening the way for a thief to break
into the building through a window or door.
And they were able to clone and manipulate RFID tags used in hotel room
key cards and corporate access cards and create a master key card to open
every room in a hotel, office or other facility. He was able, for example,
to clone Mifare, the most commonly used key-access system, designed by
Philips Electronics. To create a master key he simply needed two or three
key cards for different rooms to determine the structure of the cards. Of
the 10 different types of RFID systems he examined that were being used in
hotels, none used encryption.
Many of the card systems that did use encryption failed to change the
default key that manufacturers program into the access card system before
shipping, or they used sample keys that the manufacturer includes in
instructions sent with the cards. Grunwald and his partners created a
dictionary database of all the sample keys they found in such literature
(much of which they found accidentally published on purchasers' websites)
to conduct what's known as a dictionary attack. When attacking a new
access card system, their RFDump program would search the list until it
found the key that unlocked a card's encryption.
"I was really surprised we were able to open about 75 percent of all the
cards we collected," he says.
Attend the Black Hat Briefings and
Training, Las Vegas July 29 - August 3
2,500+ international security experts from 40 nations,
10 tracks, no vendor pitches.