Why do people say hardware wallets are unhackable?

[wendy sy]

If I insert it into a computer then the malware already on that computer can infiltrate the wallet right?

[1714961203]

1714961203

[1714961203]

1714961203

[mk4]

Nope, at least there aren't any documented cases as far as I know(as with Ledger/Trezor/Coldcard). Though I really don't recommend it (because I'm just paranoid, sometimes a bit too much), you should be able to use a hardware wallet and make a transaction using a malware-infested device because in the first place the main purpose of the hardware wallet is so the private keys won't be exposed.

The only known way you can lose funds using a hardware wallet using a malware-infested device, is when the device your using has a clipboard hijacker whereas you actually send the funds to the hacker's wallet address; but this should be easily preventable if you double check the address recipient on your hardware wallet when making the transaction.

Yes, we see a lot of hacking complaints on the Ledger and Trezor subreddit and other communities, but they're because of ignorance and carelessness of the hardware wallet user, not the fault of the hardware wallet(or the manufacturer) itself.

[toast]

I'm not expert but it is unbackable because in order for you to make a transaction you need to manually click it through the hardware wallet itself however that doesn't mean that the wallet cannot be hack, if you bought it from third party store or from just other people there's a chance that it was already modified internally so it's better if you buy it from official stores.

[mk4]

if you bought it from third party store or from just other people there's a chance that it was already modified internally so it's better if you buy it from official stores.

Most hardware wallets has software specifically made to check if hardware tampering/modifications have taken place so it's going to be unlikely, but yea I definitely wouldn't risk it even if the chances of hardware tampering bypassing the software checker are freakishly low.

[toast]

if you bought it from third party store or from just other people there's a chance that it was already modified internally so it's better if you buy it from official stores.

Most hardware wallets has software specifically made to check if hardware tampering/modifications have taken place so it's going to be unlikely, but yea I definitely wouldn't risk it even if the chances of hardware tampering bypassing the software checker are freakishly low.

Yes, I'm not aware that there's a feature that it would check itself for a hardware modifications, though it's still better to buy from official store than buy it cheaply from third party. Investing to security is much better than losing all your money.

[fearfulcalm]

I'm not specialist but i think they private key is attached with physical things, so i believe it's unhackable when we talking about software, but i think if someone understand the process and have the physical goods he can discovery the privatekey

[Ucy]

Nope, at least there aren't any documented cases as far as I know(as with Ledger/Trezor/Coldcard). Though I really don't recommend it (because I'm just paranoid, sometimes a bit too much), you should be able to use a hardware wallet and make a transaction using a malware-infested device because in the first place the main purpose of the hardware wallet is so the private keys won't be exposed.

The only known way you can lose funds using a hardware wallet using a malware-infested device, is when the device your using has a clipboard hijacker whereas you actually send the funds to the hacker's wallet address; but this should be easily preventable if you double check the address recipient on your hardware wallet when making the transaction.

Yes, we see a lot of hacking complaints on the Ledger and Trezor subreddit and other communities, but they're because of ignorance and carelessness of the hardware wallet user, not the fault of the hardware wallet(or the manufacturer) itself.

Interesting.
Wish I could see examples of such complaints to know whose fault it really is.
I always thought a good practice is to anticipate such careless users/customers and make the products as fool-proof/fraud-proof as possible.

[DannyHamilton]

I may be mistaken, but my understanding is that the physical circuits in the Ledger Nano are created such that the firmware cannot be updated without interacting with the physical buttons on the device.  Therefore, it is not possible to get malware firmware onto the device from simply plugging it into the computer.

Now, what probably IS possible is that a malicious employee (or team of employees?) at Ledger might be able to slip malware into an official Ledger Firmware update. Then users would install that firmware without knowing about the malware hidden inside.

As such, I tend to be very wary of updating the Firmware unless I know exactly why it's important, and I typically wait until a significant number of others have already updated their firmware without issue.  This doesn't guarantee security.  It's entirely possible that malware could sit silently doing nothing for days, weeks, months, or even years before it activates and causes a problem, but it at least reduces risk slightly.

Additionally, a phishing site, or other methods of manipulating people might trick someone into choosing to install malware firmware onto their device.

[mk4]

Interesting.
Wish I could see examples of such complaints to know whose fault it really is.
I always thought a good practice is to anticipate such careless users/customers and make the products as fool-proof/fraud-proof as possible.

You can only do so much to protect your customers. I own a Ledger hardware wallet myself, and it's heavily heavily implied both in the packaging and in the Ledger Live software itself that we should protect our wallet's 24-word seed and that it should only be written down. But yet, we still see people saving it on their email and such. Unfortunately those are the type of people that are simply just going to learn from mistakes.

[joniboini]

If you really want to know what kind of attacks work (while it is not practical), you can easily search on Google. There are dozens of security reports about hw wallet, and most of them require the attacker to control the HW itself. That doesn't mean all of them are good though, you should be careful with new and closed source HW.

The malware attack might happen if you use a device not created especially for HW, such as live OS on a flashdisk with Electrum imo.

[dkbit98]

If I insert it into a computer then the malware already on that computer can infiltrate the wallet right?

I don't know who said that hardware wallets are unhackable but any electronic device can potentially be hacked at some point.
If you are connecting your hardware wallet with usb cable to your computer you are not exposing your seed words or private key and there is no direct connection with internet.
There are however several attack vectors for Hardware Wallets, someone can extract your seed words if you don't have passphrase,
someone can get your paper with seed words, or you can make a mistake and enter seed words in some phishing website.

Hardware wallets are generally safer than hot wallets but only if they are open source and without some hidden code,
and like we saw in recent example when FBI was running and selling fake secure encrypted phones ANØM for 3 years and busted some criminals worldwide.
Lesson is - never trust closed source ''secure devices''.

[Pmalek]

The only known way you can lose funds using a hardware wallet using a malware-infested device, is when the device your using has a clipboard hijacker whereas you actually send the funds to the hacker's wallet address; but this should be easily preventable if you double check the address recipient on your hardware wallet when making the transaction.

There was a vulnerability discovered a while back that allowed an attacker (don't ask me how) to increase the transaction fees for bitcoin transactions involving segwit inputs. Apparently, the user wasn't aware of the fee increase. However, that 'bug' couldn't lead to funds being stolen or sent to a different address. It has also been fixed in the meantime.    

Yes, we see a lot of hacking complaints on the Ledger and Trezor subreddit and other communities, but they're because of ignorance and carelessness of the hardware wallet user, not the fault of the hardware wallet(or the manufacturer) itself.

In 99% of such cases it's the fault of the user who downloaded a fake, malware-infected software, got phished by entering his seed somewhere, or had his seed stolen from his email, Google drive or some other online service. And those people accuse the hardware wallet creators for the problems they themselves created.   

[ranochigo]

Yes, I'm not aware that there's a feature that it would check itself for a hardware modifications, though it's still better to buy from official store than buy it cheaply from third party. Investing to security is much better than losing all your money.

Correct. The device actually cannot check for hardware modifications by itself. The most that it can do is to verify that the firmware is authentic and there is nothing done to it. With the PCB, you can embed certain stuff and trick the user into thinking that the device is unmodified. That is why hardware wallets also have tamper-proof mechanism and people are strongly discouraged from purchasing used HW wallets.

With regards to OP's question, the MCU and the firmware is designed to not leak sensitive information via the USB and the bootloader should validate for the authenticity of the firmware as well. There is nothing against exploitation of vulnerabilities though, Trezor does have quite a serious vulnerability with their devices which allows seeds to be readily extracted. Hardware wallets are far more hardened than your regular computer but that doesn't mean it is immutable to exploits.

[mk4]

There was a vulnerability discovered a while back that allowed an attacker (don't ask me how) to increase the transaction fees for bitcoin transactions involving segwit inputs. Apparently, the user wasn't aware of the fee increase. However, that 'bug' couldn't lead to funds being stolen or sent to a different address. It has also been fixed in the meantime.    

If this indeed occurred, I could almost guarantee that it was due to a hardware-modified hardware wallet, not necessarily a software thing. This is one of the reasons why I think we should normalize heavily advising against buying through 3rd parties to prevent that slight chance of purchasing a modified one.

[Pmalek]

If this indeed occurred, I could almost guarantee that it was due to a hardware-modified hardware wallet, not necessarily a software thing.

I was curious about it myself, so I decided to look into it. It was indeed a software-related vulnerability, and not something to do with a modified hardware device.

They explain in the FAQ in this report that the client's application has to be compromised. You would have to be tricked into downloading a fake Ledger app. It really didn't make much sense for an attacker to play around with this thing because he can't steal any money. All he can do is make you pay much higher mining fees and he doesn't benefit from that.

[o_e_l_e_o]

The only known way you can lose funds using a hardware wallet using a malware-infested device, is when the device your using has a clipboard hijacker whereas you actually send the funds to the hacker's wallet address; but this should be easily preventable if you double check the address recipient on your hardware wallet when making the transaction.

In addition to the vulnerability discussed by Pmalek above, here is another one which requires malware on the computer and requires no physical access to your hardware wallet: https://monokh.com/posts/ledger-app-isolation-bypass

The attack essentially meant that any time you were interacting with any coin which is a fork of bitcoin (such as Litecoin, BCash, or also Bitcoin Testnet), your hardware wallet could be fooled in to accepting a bitcoin transaction disguised as an altcoin transaction. You would then confirm it, believing you are confirming an altcoin transaction, resulting in the loss of your bitcoin. It has now been patched.

The underlying message is that there are almost certainly other ways for malware to attempt to steal coins from an attached hardware wallet without the hardware wallet itself being tampered with or an attacker having physical access. User error will always be the biggest risk with using pretty much any wallet (particularly when it comes to backing up seed phrases), and attacks such as this one will be incredibly rare, but I wouldn't just go plugging your hardware wallet in to any malware ridden computer and assume you will always be 100% safe.

[Kakmakr]

What is the Malware going to do? I have to physically "push" a button on my device to confirm the transaction, so the Malware has to intercept that input and then duplicate that to the computer? I also use the buttons on the device to enter my password.. so the Malware cannot "capture" the keyboard.. because I am not using my keyboard to login to the device. 

There was someone that used specialized tools to physically hack the device (Trezor) by intercepting the signal, but they had access to the device and they did not do this over the Internet. (Long time ago..when the Trezor was just launched) 

[DannyHamilton]

What is the Malware going to do?

For one thing, the Malware could update the firmware on the device to change it's behavior.

Additionally, it could send a transaction to the device to be signed which is modified in such a way that you fail to notice it isn't the transaction that you intended to send.  Then, since you don't notice it, you push the buttons to sign the transaction.

[NeuroticFish]

What is the Malware going to do?

For one thing, the Malware could update the firmware on the device to change it's behavior.

Additionally, it could send a transaction to the device to be signed which is modified in such a way that you fail to notice it isn't the transaction that you intended to send.  Then, since you don't notice it, you push the buttons to sign the transaction.

That's correct. Did you notice that the HW doesn't show your change address to verify it? Now just imagine that the HW is made to work as a certain external (hardcoded?) address is the change address. You won't notice a thing and your money will go *spoof*

[ranochigo]

That's correct. Did you notice that the HW doesn't show your change address to verify it? Now just imagine that the HW is made to work as a certain external (hardcoded?) address is the change address. You won't notice a thing and your money will go *spoof*

The firmware should recognize the change address and that is why some HW wallets don't display it. Specifically, ColdCard checks for the change address with their internal list. I can't speak for the other wallets as I don't use them.

If the firmware is modified in a way that changes the change address, it could've tricked you in many other ways as well.