We updated our RSA SSH host key
github.blog
github.blog
Github's customers trust GH/MS with their code, which, for most businesses, is a high value asset. It wouldn't surprise me if this all results in a massive lawsuit. Not just against GH as a company, but also to those involved (like the CISO). Also, how on earth was it never discovered during an audit that the SSH private key was plain text and accessible? How has GH ever been able to become ISO certified last year [0], when they didn't even place their keys in a HSM?
Obviously, as a paying customer I am quite angry with GH right now. So I might be overreacting when I write this, but IMO the responsible persons (not talking about the poor dev that pushed the key, but the managers, CISO, auditors, etc.) should be fined, and/or lose their license.
[0] https://github.blog/2022-05-16-github-achieves-iso-iec-27001...
Regarding the secrecy of the data you host at github, you should operate under the assumption that a sizeable number of github employees will have access to it. You should assume that it's all sitting unencrypted on several different disk platters replicated at several different geographically separated locations. Because it is.
One of the things that you give up when you host your private data on the cloud is controlling who and under what circumstances people can view or modify your private data. If the content of the data is enough to sink you/your company without remedy you should not store it on the cloud.
When we signed on with GH as a paying customer over a decade ago, they were quite clear that private repos should not be considered secure storage for secrets. It’s not encrypted at rest, and GitHub staff have access to it. It takes only a few clicks to go from private to public.
[1] Which is sort of tautological and silly, because that's true of all sites and all host keys. What the comment was trying to imply was that the choice of storage of this key invalidates any trust we might have in GitHub/Microsoft regarding key management, and that therefore we shouldn't trust them. Which is also tautological and silly. Either you trust them or you don't, that's not a technological argument.
There are plenty of tools to either keep them encrypted (we just use simple GPG, but our team is small) or just auto-generate and never show to user in the first place (various key vaults that can be used from automation like HashiCorp's Vault)
If ssh token based auth is used, it's different of course, because then the server gets access to the token. Ideally Github would invalidate all their auth tokens as well.
The fun fact is that a token compromise (or any other attack) can still happen any point in the future with devices that still have outdated ssh keys. That's a bit unfortunate as no revocation mechanism exists for ssh keys... ideally clients would blacklist the ssh key, given the huge importance of github.
Well, at least not without encryption that is under your control.
ISO/IEC 27001:2013 doesn’t say you have to store private keys in HSMs? It just requires you to have a standards compliant ISMS that implements all Annex A controls and all of the clauses. Annex A and the clauses don’t specifically mandate this.
If you can convince an auditor that you have controls in-place that meet the standard for protecting cryptographic media you basically meet the standard. The controls can be a wide variety of options and don’t specifically mandate technical implementation details for many, many things.
You shouldn’t rely on ISO/IEC 27001:2013 as attestation for technical implementation details here. Just because your auditor would red flag you doesn’t mean all auditors would. The standard is only as effective as the weakest, cheapest auditor, and there are perverse incentives that make auditors financially incentivized to certify companies due to recurring revenue.
Thanks for the insight, good advice.
But also from the same GH article [0]:
The ISO 27001 certification is the latest addition to GitHub’s compliance portfolio, preceded by SOC and ISAE reports, FedRAMP Tailored LiSaaS ATO, and the Cloud Security Alliance CAIQ.
Do you have any knowledge on one of these certifications (for exmaple FedRAMP) that puts any restrictions on handling key material?
[0] https://github.blog/2022-05-16-github-achieves-iso-iec-27001...
Not to diminish the problems of having a large entity like Github handle a private key like that, but if that was your level of paranoia, you probably should have used commit signatures all along and not relied on Github to do that job for you.
I also don't want to diminish the concerns around Github or similar orgs losing control of a private key, but the far more realistic concern for the vast majority of threat models is often put to the wayside in favor of what amounts to a scary story. Rather than the straightforward key removal and replacement that this should be, I (and surely many others) have spent all morning combatting this specific FUD that cropped up on HN with leadership and many engineers. It's actually quite detrimental to quickly remediating the actual concerns introduced by this leak.
I understand that security inspires people to be as pedantic as possible - that's where some big exploits come from on occasion - but I really hope the average HN narrative changes toward "what is your actual, real-world threat model" vs. "here is a highly theoretical edge-case scenario, applicable to very few, that I'll state as a general fact so everyone will now wonder if they should spend months auditing their codebase and secrets". Put simply: this is why people just start ignoring security measures in the real world. Surely someone has already coined the term "security fatigue".
It's all just a bit unbalanced, and definitely becomes frustrating when those suggesting these "world is burning" scenarios didn't even take the available precautions that apparently would satisfy their threat model (i.e. commit sigs, as you suggested)
Ok, end rant :)
Requiring all commits to be signed by trusted keys avoids the risks associated with someone tampering with a repository hosted on GitHub if they are able to get access to it, although it doesn't protect against code being leaked.
See here for details: https://git-scm.com/book/en/v2/Git-Tools-Signing-Your-Work
Many signed commits ("Verified" in green) on GitHub are signed with GitHub's own private key¹, not the committer's key. There's a good technical reason, but many committers don't realise their own signing key isn't the one used on their signed commits to the main branch.
That GitHub private key would be a fun one to have leaked!
It would invalidate most of the "Verified" flags on commits on most repo main branch histories.
(¹ GitHub uses GitHub's own commit-signing key when you use the GitHub GUI to merge or squash-merge commits if you already signed them, and the resulting commits show as "Verified" the same as conmits signed by your own key. So you do have to sign with your own key,, but it's not the key used on the main branch commits everyone sees whrn they visit and doenload. Many controlled workflows require approved PRs to be merged to main/master through GitHub's tools, and many users default to using the GUI for PR merges by choice.)
For a host key? Like I get that being able to impersonate Github isn't great as far as state level actors having the ability to do this but you do know the actual transport layer keys are ephemeral and aren't derived at all from the host key, right?
Not just nation state actors, but basically anyone in a position to MITM.
Also, you don't have to be a nation state actor to extort a GH employee. Any bad guy can do a "give me this key or I'll hurt your kid". People are being extorted for a lot less.
There are billions of dollars of assets flowing through GH's infrastructure, for the sake of safety (!= security) of Github's employees, nobody should ever have access to key material.
The host key is the only thing ensuring you’re actually talking to GitHub.com when you push code.
To add to sibling comments, it should not have been possible to make this mistake. That it was possible is concerning.
Great! Then I can communicate confidentially with whomever is MITM'ing me.
/s
Number of people just not blindly using TOFU with github over ssh must be quite low.
Who here went here https://docs.github.com/en/authentication/keeping-your-accou... and added the keys manually to known_hosts before using github for the first time?
Do you realize that the code just sits on GitHub servers even if it's private?
If you have any degree of paranoia, why do you put your code into GitHub?!?!
Like, if you work on code which
ISO 27001 certification does not require you to put keys into an HSM. The standard requires you to have controls in place, be aware of your risks and to maintain a risk register. But in no way does the standard require HSM's.
The standard would even be OK with storing this on a floppy drive if the risks surrounding that were identified and mitigated (or accepted).
In fact, this is not a supported option in openssh.
You probably also never met a single person where the SSH interface sees millions of sessions as day with valuable assets (code) being transported over said sessions.
> In fact, this is not a supported option in openssh.
This definitely is supported. Though documentation for this is often HSM vendor specific, which if heavily NDA'd. So that's why you probably haven't found much information about it.
It looks like they currently use the `ssh_bind_options_set` function with `SSH_BIND_OPTIONS_HOSTKEY` to set the host keys which means they exist on disk at some point. HSM aside, I believe it would be possible to use the `ssh_bind_set_key` and deserialize them from a secret vault so they only exist in the memory of the ssh daemon.
Obviously they also just straight up have enough resources to fork the code and modify it to use an HSM.
Source: looking at their ssh server portion of `babeld` in ghidra right now as part of hunting for bug bounties.
https://framkant.org/2017/10/strong-authentication-openssh-h...
I suppose (Open?)SSH's PKI mode could support a model like that, but as others have noted here, this requires much more manual work on the user's side than comparing a TOFU key hash.
Maybe that model could be extended to allow TOFU for CAs, though? But I think PKI/CA mode is an OpenSSH extension to the SSH protocol as it is, and that would be a further extension to that extension...
Storing the key in some kind of credential vault that can only be accessed from the hosts that need it at startup would usually be enough to prevent this particular kind of error (unless you're giving root on those boxes to people without enough sense to avoid checking private keys into git, in which case you've probably got worse problems).
By no means do I want to see the dev get fined or blackballed from the industry.
But if there is any 1 person responsible, it’s the person who did it. The reason why the dev shouldn’t be fined/blackballed is because it’s not just 1 person’s fault. I mean, fining or booting his manager out of the industry? Really?
1) Nation state level actors can probably insert or compromise high level staff, or multiple high level staff, at any given company, and perform MITM attacks fairly easily. And some could compel turning over your code or secrets more directly anyway. Not worth worrying about this scenario: nobody working on anything truly sensitive should be using any externally hosted (or even externally networked) repositories.
2) It is much more difficult for other actors to do a MITM attack, and if they did, they'd probably have access to your code more directly.
3) Your code actually isn't worth much to anybody else. Imagine someone launching a complete clone of HN or any other site. Who cares? Nobody. What makes your code valuable is that you have it, and that you have a network and relationship with your customers. If somebody stole my company's codebases, I'd feel sorry for them, that they are going to waste any time wading through these useless piles of shit. The only potential problems are if secrets or major vulnerabilities are exposed and provide a path for exploit (like ability to access servers, services, exposing potential ransomware attacks).
This leak opened a time window big enough for that to happen. We may or may not know if it did. I doubt this info would be offered to the public because it would sink the business.
I think this suggests we need more information from github. For instance GH employees may not always have had live access to this key, this could have happened as part of an operation that gave temporary access to an employee only recently. Or it could have been stored plaintext on multiple employees' home computers since creation.
When was the leaked key created anyway?
The cloud is always the convenience of someone else’s computer over some amount of security.
"We have no reason to believe that the exposed key was abused, but out of an abundance of caution, we are going to expose 50 million users to a potential MITM attack unless they are extremely careful."
Not a single word in the post about whether this impact was even considered when making the decision to update the key. Just swap the key, and fuck the consequences. Same with the mass password resets after a compromise that some services have done in the past years. Each of those is any phishing operation's dream come true.
So MITM for some of 50m users is strictly better than MITM for all of 50m users.
> leaked in public repo
Me: Yeah, that's why they are doing it.
How do you know this?
Github runs scanner for private keys in public and private repos and notifies owner (I did it once so I know ... ;)). So some Github engineer likely would have received such an email if what you say is true. Hilarious.
If the key was breached and Github just didn't know it, then a breach happened, then only Github would be to blame.
If Github rotates its key, and somebody suffers a MITM attack, the blame is more diffuse. Why didn't they verify the key out of band?
SHA256:uNiVztksCsDhcc0u9e8BujQXVUpKZIDTMczCvj3tD2s (RSA)
SHA256:br9IjFspm1vxR3iA35FWE+4VTyz1hYVLIE2t1/CeyWQ (DSA - deprecated)
SHA256:p2QAMXNIC1TJYWeIOttrVc98/R1BUFWu3/LiyKgUfQM (ECDSA)
SHA256:+DiY3wvvV6TuJJhbpZisF/zLDA0zPMSvHdkr4UvCOqU (Ed25519)We humans really aren't cut out for this, are we.
Not all of us are familiar enough with the SSH protocol to understand how to "double check the expected value"? Where can I determine what the expected value should be?
It should error like this:
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
@ WARNING: REMOTE HOST IDENTIFICATION HAS CHANGED! @
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
IT IS POSSIBLE THAT SOMEONE IS DOING SOMETHING NASTY!
Someone could be eavesdropping on you right now (man-in-the-middle attack)!
It is also possible that a host key has just been changed.
The fingerprint for the RSA key sent by the remote host is
SHA256:uNiVztksCsDhcc0u9e8BujQXVUpKZIDTMczCvj3tD2s.
Please contact your system administrator.
https://docs.github.com/en/authentication/keeping-your-accou...
Yes, this assumes the github-hosted docs and your SSL connection to them are not also compromised, but it's far better than not checking at all.
Or the parts below it about updating and verifying in other ways.
What do I have to do to ensure I connected to the right server?? I thought just making sure the correct RSA key was in known_hosts would be enough?
If you read the blog post on a verified domain and saw the new key and updated manually, or you deleted the known key and verified the key fingerprint when it warned you about an unknown key, you should be good to go. Here, you trust the people who issue TLS certificates and you trust github to be in control of their domain name, so you can be reasonably confident that the key they advertised on their website is the correct key. If your internet connection was compromised, you would have got an error message when you connected to https://github.blog (because they wouldn't have a certificate from a trusted certificate issuer) or when you connected to the git server (because they wouldn't have the key you just trusted).
If you saw the blog post and then removed the old key and told ssh to save the new key it's receiving without checking it matches the value on the webpage, you might have a problem. The connection to github's ssl could have been compromised, and if you just accepted whatever it told you, you have no trusted intermediate to verify that the key is trustworthy. All you know is that each time you connect to github's server hereafter, you're either connecting to a server with the same key (no error), or you're connecting to one that doesn't have the same key (error message). But whether you can trust that key? You don't know that. You just know it's the same key.
But even if you did the latter, all is not lost. You can look at the known_hosts file (on Linux and MacOS it's ~/.ssh/known_hosts) and check the fingerprint. If it's what they advertise, then you're good to go. If it's different, you should fix it and find people who can help you deal with a security incident.
The reason people are raising a flag is that today, lots of people will be rotating their key. That means if you're looking to target someone, today is the day to do it. Even if 90% of people do it the proper way, by manually verifying the key, that still means there's going to be a lot of people who could be victimised today.
I assume it's safe because the SSL cert for docs.github.com is probably not compromised, so it's giving us the right key, and compromising docs.github.com would be extra effort and is unlikely to happen.
However, I wonder what kind of steps an MITM attack would have to perform, I assume one of the easiest would be compromising my local DNS server, since regular DNS does not offer a high level of security, then github.com resolves to the attacker's IP and the attack works. Do you have examples of such attacks that don't involve a virus already active on the end user's PC? Maybe if someone owns an IP previously owned by Github that is still somehow advertised as being Github by some DNS lagging behing?
The best practice is to verify the fingerprint out of band using a secure channel. In this case, that's HTTPS and docs.github.com. If (hypothetically) docs.github.com was also compromised, then you don't have a secure channel.
https://en.m.wikipedia.org/wiki/Man-in-the-middle_attack has some MITM examples.
We should have a blockchain for certificates btw. That would be such an amazing solution to this problem. You could advertise ahead of the time that you are changing certificates and we could verify that it was in fact you.
It was trusted to clone code into the infrastructure of hundreds of thousands of organizations. People pin it everywhere.
With this key someone could infect the infrastructure of fintech companies and steal billions of dollars. I know this well because I run a security consulting company focusing mostly on that industry. Mainly this is possible because almost no companies check for git commit signing, and Github does not enforce it anyway, and I digress.
This key held enough power over value that some might have used violence to acquire it.
With that context of course they chose to place this key in a hardware security module controlled with an m-of-n quorum of engineers to ensure no single human can steal it, expose it, or leak it. Right? ... right?
Nope. They admit they just stuck it in a git repo in plain text where any engineer could have copied it to a compromised workstation, or been bribed for it, for who knows how many years. Worse, it was not even some separate employee only intranet git repo, but their own regular public production infra and someone had the power to accidentally make it public.
I have no words for this level of gross negligence.
Maybe, just maybe, centralizing all trust and security for most of the worlds software development to a proprietary for-profit company with an abysmal security reputation was not the best plan.
I will just leave this here: https://sfconservancy.org/blog/2022/jun/30/give-up-github-la...
In reality, the vast majority of users don't pay attention to SSH host keys at all.
Even if an attacker got hold of this private host key, they'd have to be able to MitM the connections of their target.
Next, they have to decide what they want to do.
If they want to send malicious code to someone doing a 'git pull', they'd have to craft the payload specifically for the repo being pulled from. Not impossible, but difficult.
If they want to "steal" source code from someone doing 'git push' (perhaps to a private repo on GitHub), that's a bit easier, as they can just tell the client "I have no objects yet", and then the client will send the entire repo.
And, again, they'd have to have the ability to MitM some git user's connections. Regardless, there is no way that they could change any code on github.com; this key would not give them any access to GH's services that they don't already have.
So I think your anger here is a bit over the top and unnecessary.
I agree that it's pretty bad that some GH employee even had the ability to get hold of this private key (sure, probably should be in an HSM, but I'm not particularly surprised it's not) in order to accidentally leak it, but... shit happens. They admitted the mistake and rotated the key, even though it's likely that there was zero impact from all this.
Imagine you control the right router on a company wifi, or any home wifi a production engineer works from and suddenly you can cause them to clone the wrong git submodule, the wrong go package, or the wrong terraform config.
If you knew a CI/CD system blindly clones and deploys git repos to prod without signature checks, and that prod is a top 10 crypto exchange with 1b of liquidity in hot wallets, then suddenly a BGP attack to redirect DNS is a good investment. Myetherwallet got taken over for 15 minutes with a BGP so this is not hypothetical.
Should that be the case? Of course not. But the reality is I find this in almost all of the security audits I do for fintech companies. Blind trust in Github host keys is industry standard all the way to prod.
This is not hard. If it were hard, we wouldn't need encrypted connections.
If I were a nation state, it would be trivial to position an attack at the edge of a network peering point for a given service. How do I know? Our own government did it for 10+ years.
Cyber criminals often have the same tactics and skills and can find significant monetary reasons to assume heightened risk in order to pull off a compromise.
Random black hats enjoy the challenge of compromising high value targets and can find numerous ways to creatively infiltrate networks to perform additional attacks.
Even without gaining access to a network, virtually anyone on the internet can simply push a bad BGP config and capture traffic for an arbitrary target. Weird routing routinely happens that nobody can definitely say isn't such an attack.
Genuinely asking because I’ve struggled with this question.
1. You compile a deterministic unikernel appliance-style linux kernel with a bare bones init system
2. You deploy it to a system that supports remote attestation like a nitro enclave.
3. It boots and generates a random ephemeral key
4. m-of-n engineers compile the image themselves, get the same hash, and verify the remote attestation proof confirming the system is running the bit-for-bit trusted image
5. m-of-n engineers encrypt and submit shamirs secret shares of the really important private key that needs protecting
6. key is reconstituted in memory of enclave and can start taking requests
7. Traffic goes up and autoscaling is triggered
8. New system boots with an identical account, role, and boot image to the first manually provisioned enclave
9. First enclave (with hot key) remotely attests the new enclave and obtains its ephemeral key (with help of an internet connected coordinator)
10. First enclave encrypts hot key to new autoscaled enclave
11. rinse/repeat
This is unfortunately not how SSH works. It needs to be unlocked for every incoming connection.
You raise valid hypotheticals about the security of the service... but fixing it involves removing SSH host key verification from Github; better OpsSec would not fully resolve this issue.
Even if they wanted the most quick and dirty lazy option with no policy management, they could stick the key in a PKCS#11 supporting enclave every cloud provider supports these days. OpenSSH natively supports them today.
At a minimum they could have placed their whole ssh connection termination engine in a immutable read only and remotely attestable system like a Nitro enclave or other TEE. You do not need to invent anything new for this.
There are just no excuses here for a company with their size and resources, because I do this stuff all the time as just one guy.
Yep. Well, certificates exist exactly to bridge the GP's requirement with your reality.
You could also use things like Nitro enclaves which have all the same specs as a regular EC2 instance.
Tons of options. They clearly chose none of them.
These kinds of changes suuuuuuck. Messing with known_hosts file is not always something easy to do. Might require a image rebuild, if you have access at all.
Do not put long lived cryptographic key material in the memory of an internet connected system. Ever. It is a really easy to understand rule.
- When exactly was the key leaked? Yesterday? A month ago?
- How long was it visible for? And no, "briefly" doesn't cut it.
- Did any internet traffic reach the page while it was exposed? We know you log everything, so it is a yes or no answer.
If any of these answers were pretty, I imagine they would have explicitly included them in the post.
I don't know how long exactly, but in theory you can subscribe to a stream of ALL events happening at GitHub by fetching from this endpoint: https://api.github.com/events
With these events you know what new repositories are created and what changes are pushed, so you can fetch every new change. Once something gets pushed to a public repository, it's very likely that some spider will have fetched it within a few minutes.
Wow I am shocked that they allow "firehose" access not only for free, but without even an API key.
Given enough disk and bandwidth, does this mean you could keep your own copy of all of github? I'd love to be able to grep the whole thing.
> We have no reason to believe that the exposed key was abused and took this action out of an abundance of caution.
This is not verifiable, right? As the authentication method has no public and required revocation source, and given that the key, if having leaked, likely will be acquired by an authoritarian government, they can selectively MITM organizations and users that are not aware of this blog post.
> This week, we discovered...
So they found out about it sometime between Monday and Thursday, and rotated it Thursday evening.
"This week, we discovered that GitHub.com’s RSA SSH private key was briefly exposed in a public GitHub repository"
(sorry, the comments are mangling this, clean version at https://gist.github.com/robbat2/b456f09b7799f4dafe24115095b8...)
``` # You might need to insert this in a slightly different place cat >>/etc/ssh/ssh_config <<EOF Host * RevokedHostKeys /etc/ssh/ssh_revoked_hosts EOF
cat >>/etc/ssh/ssh_revoked_hosts <<EOF # https://github.blog/2023-03-23-we-updated-our-rsa-ssh-host-k... ssh-rsa AAAAB3NzaC1yc2EAAAABIwAAAQEAq2A7hRGmdnm9tUDbO9IDSwBK6TbQa+PXYPCPy6rbTrTtw7PHkccKrpp0yVhp5HdEIcKr6pLlVDBfOLX9QUsyCOV0wzfjIJNlGEYsdlLJizHhbn2mUjvSAHQqZETYP81eFzLQNnPHt4EVVUh7VfDESU84KezmD5QlWpXLmvU31/yMf+Se8xhHTvKSCZIFImWwoG6mbUoWf9nzpIoaSjB+weqqUUmpaaasXVal72J+UX2B+2RPW3RcT0eOzQgqlJL3RKrTJvdsjE3JEAvGq3lGHSZXy28G3skua2SmVi/w4yCE6gbODqnTWlg7+wC604ydGXA8VJiS5ap43JXiUFFAaQ== EOF ```
$ ssh-keygen -lf github.old.pub
2048 SHA256:nThbg6kXUpJWGl7E1IGOCspRomTxdCARLviKw6E5SY8 no comment (RSA)
https://web.archive.org/web/20230320230907/https://docs.gith...
(please check my work on your own machines and don't take my attestation on faith!)
I expanded on your gist:
https://gist.github.com/e12e/0c1868479c0b8d0a52914d44be66d76...
Thanks for the gist though, seems helpful!
RevokedHostKeys doesn't accept ~ for your home directory... while things like ControlPath will.
I'd rather confine this to my account, but I either have to use a relative path that doesn't always work... or a fully qualified path that includes my username (and may change)
> ... out of an abundance of caution, we replaced our RSA SSH host key used to secure Git operations for GitHub.com
Yeah, that's not an "abundance of caution." That's the bare minimum response at that point. What's the "not cautious approach?" Make the repo private and go on your merry way?
I suppose "abundance of caution" might apply, if they determined that the only ways it could've leaked were from requests that were logged, and they've removed all the ways and checked all the logs.
But if I had to guess, even a brief exposure can be picked up by bots (and perhaps they can already see log entries for this). Even if no one at all picked it up, there'd still be the question of whether traces of it are still left behind on various infrastructure, in storage and caches (even ML training?) not intended for key safety.
github.com ssh-rsa AAAAB3NzaC1yc2EAAAABIwAAAQEAq2A7hRGmdnm9tUDbO9IDSwBK6TbQa+PXYPCPy6rbTrTtw7PHkccKrpp0yVhp5HdEIcKr6pLlVDBfOLX9QUsyCOV0wzfjIJNlGEYsdlLJizHhbn2mUjvSAHQqZETYP81eFzLQNnPHt4EVVUh7VfDESU84KezmD5QlWpXLmvU31/yMf+Se8xhHTvKSCZIFImWwoG6mbUoWf9nzpIoaSjB+weqqUUmpaaasXVal72J+UX2B+2RPW3RcT0eOzQgqlJL3RKrTJvdsjE3JEAvGq3lGHSZXy28G3skua2SmVi/w4yCE6gbODqnTWlg7+wC604ydGXA8VJiS5ap43JXiUFFAaQ==
[1] https://github.blog/changelog/2022-01-18-githubs-ssh-host-ke...
I looked at my `~/.ssh/known_hosts` file and that key is associated with a few IP addresses in addition to github.com. Those lines stayed after I ran `ssh-keygen -R github.com`.
I imagine that I also need to remove those other lines manually, but isn't that something that GitHub should have mentioned? I'm not sure in which circumstances these got added either…
- 192.30.253.112
- 140.82.114.3
|1|{base64-encoded-string?}|{another-base64-encoded} ssh-rsa AAA{long-string}
Unfortunately there's no way to make this easy. To associate a CA key with a specific domain you need to add something like:
@cert-authority github.com (key)
to known_hosts, and this (as far as I know) can't currently be managed in the same way that keys are generally added to known_hosts on first use when you ssh to something. So instead we're left with every github.com SSH frontend using the same private keys, and those can't be in an HSM because it simply wouldn't scale to the required load. Which means the private key exists in a form where this kind of thing can occur.
HTTPS has the same problem. Compromise any frontend, and you can MITM traffic to any other front-end till the expiry of the certificate (usually 90 days).
I would really like to see some kind of almost realtime delegation - so that for example, every second a new sub certificate could be generated and signed by a HSM, valid for only 2 seconds.
With HTTPS certs you usually have at most 90 days of impact when a key is leaked (less if you revoke and software is checking CRL). GitHub used the same RSA key for over a decade, they may have continued using this key for quite some time more had they not noticed the leak this week.
My immediate assumption wasn't that GitHub had changed keys. I thought that my computer was having some sort of problem. Only after searching around for a bit did I find out that yes, GitHub changed their private key.
Couldn't they have e-mailed all of their users and/or customers, to tell us what had happened?
• When you run a git command and git tells you about the mismatching key, it will prompt you to enter `y` or `n` but not actually let you do so.
• You need to run `plink github.com` to get the same prompt. You can then enter `y` to accept it.
• You should only accept it if it says the new fingerprint is d5:2c:63:d9:bc:75:9d:de:b1:4e:36:28:9f:7a:9c:39. If it is any different, you may be targeted by an attacker.
• The new key is stored in the file `C:\Users\<your name>\.ssh\known_hosts`. The new key at the time of this writing is `github.com ssh-rsa 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`.
• In the future, if the key ever changes again, you can look it up on GitHub’s SSH key fingerprints page [0].
[0] https://docs.github.com/en/authentication/keeping-your-accou...
Seriously? How that happened is deeply concerning.
And why weren't the other keys exposed?
$ ssh -o VerifyHostKeyDNS=yes github.com
The authenticity of host 'github.com (140.82.121.3)' can't be established.
ED25519 key fingerprint is SHA256:+DiY3wvvV6TuJJhbpZisF/zLDA0zPMSvHdkr4UvCOqU.
No matching host key fingerprint found in DNS.
What a security nightmare…
# Add this to the top of ~/.ssh/config
RevokedHostKeys /home/username/.ssh/revoked_host_keys
cd ~/.ssh
mkdir revoked_host_keys.d
echo 'ssh-rsa AAAAB3NzaC1yc2EAAAABIwAAAQEAq2A7hRGmdnm9tUDbO9IDSwBK6TbQa+PXYPCPy6rbTrTtw7PHkccKrpp0yVhp5HdEIcKr6pLlVDBfOLX9QUsyCOV0wzfjIJNlGEYsdlLJizHhbn2mUjvSAHQqZETYP81eFzLQNnPHt4EVVUh7VfDESU84KezmD5QlWpXLmvU31/yMf+Se8xhHTvKSCZIFImWwoG6mbUoWf9nzpIoaSjB+weqqUUmpaaasXVal72J+UX2B+2RPW3RcT0eOzQgqlJL3RKrTJvdsjE3JEAvGq3lGHSZXy28G3skua2SmVi/w4yCE6gbODqnTWlg7+wC604ydGXA8VJiS5ap43JXiUFFAaQ==' > revoked_host_keys.d/github-leak-2023.03.23
ssh-keygen -k -f revoked_host_keys revoked_host_keys.d/*
ssh-keygen -l -F github.com
[1] https://docs.github.com/en/authentication/keeping-your-accou...
[2] https://docs.github.com/en/authentication/keeping-your-accou...
It still shouldn’t have been easily accessible to anyone except the instances running the SSH service.
Yes, it makes things hard and unconventional to set up. But GitHub is not some small website.
Github is host to a large percent of US tech IP. Pretty concerning if you extrapolate.
When I tried to pull from my repo I got the warning message, right
I removed the old keys with ssh-keygen -R github.com
Then, trying with `ssh -T git@github.comp` I see this
The authenticity of host 'github.com (140.82.121.3)' can't be established. ECDSA key fingerprint is SHA256:p2QAMXNIC1TJYWeIOttrVc98/R1BUFWu3/LiyKgUfQM.
So, first thing of course, is that the fingerprint does not match the one in the blogpost which is SHA256:uNiVztksCsDhcc0u9e8BujQXVUpKZIDTMczCvj3tD2s
Second problem, is that ip 140.82.121.3 seems to be reported as HIGH RISK[0]
So basically, how should I proceed? I am not security expert but I would say I am not an illiterate on this, and I have no idea. I guess the majority of users would you accept the new key, but is this the right move? I would need to do it if I want to do some work, that is for sure
EDIT: Formatting
[0]https://www.ipqualityscore.com/free-ip-lookup-proxy-vpn-test...
I get this error each time I interact with Github: The authenticity of host 'github.com (140.82.112.3)' can't be established. ECDSA key fingerprint is SHA256:p2QAMXNIC1TJYWeIOttrVc98/R1BUFWu3/LiyKgUfQM.
When I type 'yes' I get this added to my host file: 楧桴扵挮浯ㄬ〴㠮⸲ㄱ⸲″捥獤ⵡ桳㉡渭獩灴㔲‶䅁䅁㉅橖䡚桎塌潎呙瑉浢穬䡤祁呎䅙䅁䥁浢穬䡤祁呎䅙䅁䉂䕂䭭䕓橎䕑穥浏歸䵚㝹灯杋䙷㥂歮㕴剙奲橍畎㕇㡎男杒㙧䱃扲㕯䅷呤礯瘶洰噋唰眲地㉚䉙⬯含潰正㵧
I manually added the new RSA SSH public key entry to my known_hosts file (like they say in the blogpost)
Then ran ssh -T git@github.com and got
Warning: Permanently added the RSA host key for IP address '140.82.121.3' to the list of known hosts. Hi gassius! You've successfully authenticated, but GitHub does not provide shell access.
Then, when trying a git pull, I got this:
Warning: the RSA host key for 'github.com' differs from the key for the IP address '140.82.121.4' Offending key for IP in ~/.ssh/known_hosts:63 Matching host key in ~/.ssh/known_hosts:64
So basically, the Offending key is the one I added manually as per blogpost?
Ok, I am in Europe, and this seems like an issue of global distribution network or something, but this is not great AT ALL, either the blogpost information is not complete or something fishy is going on
UPDATE: The replies makes clear what I was seeing those errors and make sense. Thanks
EDIT: Formatting and Acknowledge of the situation per replies
Not in the blog post but the blog post points to my link which is the official documentation.
1. https://docs.github.com/en/authentication/keeping-your-accou...
ssh-keygen -R github.com
ssh-keygen -R 140.82.121.3
Warning: the RSA host key for 'github.com' differs from the key for the IP address '140.82.121.3'
Offending key for IP in /home/tom/.ssh/known_hosts:77
Matching host key in /home/tom/.ssh/known_hosts:102 sed -i.github-removed '/AAAAB3NzaC1yc2EAAAABIwAAAQEAq2A7hRGmdnm9tUDbO9IDSwBK6TbQa/d' ~/.ssh/known_hosts> GitHub.com’s RSA SSH private key was briefly exposed in a public GitHub repository
What the... That's not "an abundance of caution". That's the only possible course of action.
Do they expect people to think "Wow, Github leaked a key, but even without knowing if anyone snagged it, they're still replacing it. Wow, they go above and beyond."
It's so ridiculous.
After all it’s them hosting and serving the requests for that (and every other) repo.
Only _reasonable_ course of action. Possible is to do nothing =)
That's the "caution" part.
https://docs.github.com/en/authentication/connecting-to-gith...
(it would actually be best to update it to the Ed25519 public key, as it's the recommended algorithm when generating a new SSH key).
https://lwn.net/Articles/637156/
I'm honestly not familiar with anyone actually using host-key rotation?
I mean it seems like its clearly a key that wasn't in an HSM.. and over the lifetime, hundreds? Thousands of Github employees could of accessed it?
The current rotation allows anyone to try to fish the lazy users (like me probably) who will just trust on first use. Probably a bigger risk than key compromise, since they have logs.
It could be a better idea to use Host-key rotation, despite it making the life of a key-thief a bit easier. Just because it exposes people less against opportunistic impersonators.
2. It was only added in OpenSSH 6.8, so it missed Ubuntu 14.04 release, and only really turned up in 16.04 LTS that way, plenty of old systems it wouldn't work on.
As other posters noted, a bad actor could rotate the key to their chosen keys just as easily as GitHub could cause the rotation.
It seems like at least a `known_hosts` compromise would be "self-healing" after connecting to the legitimate github.com server once.
=====
After I followed the instructions to remove the old RSA key, `git pull` started using the ECDSA key, and now shows this warning:
Warning: the ECDSA host key for 'github.com' differs from the key for the IP address '20.205.243.166'
Offending key for IP in /home/forge/.ssh/known_hosts:87
Matching host key in /home/forge/.ssh/known_hosts:88
Are you sure you want to continue connecting (yes/no)?
ssh-keygen -f ~/.ssh/known_hosts -R 20.205.243.166
I'm not sure why the ECDSA key of the github.com IP has supposedly changed - maybe someone can clarify this?
Since GitHub's IP address is not stable, I suggest disabling the IP-checking feature using
Host github.com
CheckHostIP no
It would be so much better if standard practice was to generate and store the private key on a smartcard or the TPM, so that the only file a clueless/careless developer could upload would be a stub.
It really should be named `id_rsa.private` to help a busy developer realize they have the wrong file.
Yep. Especially given that basically all modern laptops (and some PCs) ship with TPMs and ssh can use it via the TPM PKCS#11 lib. I'm using that daily on multiple machines and it's working great.
It wasn't even that long ago that Github changed the checksums of git tarballs which also broke builds everywhere. All we got was a "oh btw we changed the archive behavior" blog post after the change went live [0].
[0] https://github.blog/changelog/2023-01-30-git-archive-checksu...
I'm upset they haven't sent out an email to every user. This affects their entire usebase and basically everyone needs to take manual action to continue working. If you haven't seen the news then you'll see failed builds/git commands.
At least put your key to https://github.com/.well-known/ssh/ed25519.pub so I don't need to Google it... And may be some day ssh will support it natively. Someone need to act first.
Here's the list of keys if anyone needs:
https://docs.github.com/en/authentication/keeping-your-accou...
> Host key for github.com has changed and you have requested strict checking.
> Host key verification failed.
It is also not mentioned on https://www.githubstatus.com/ aka status.github.com.
curl -s https://api.github.com/meta | jq -r '.ssh_keys | join("\n")' | sed 's/^/github.com /' > /etc/github_known_hosts
GIT_SSH_COMMAND="ssh -o UserKnownHostsFile=/etc/github_known_hosts" git pull
$ openssl s_client -showcerts -connect github.blog:443
CONNECTED(00000006)
depth=2 C = US, O = Internet Security Research Group, CN =
ISRG Root X1
verify return:1
depth=1 C = US, O = Let's Encrypt, CN = R3
verify return:1
depth=0 CN = github.blog
verify return:1
---
Certificate chain
0 s:CN = github.blog
i:C = US, O = Let's Encrypt, CN = R3
a:PKEY: id-ecPublicKey, 256 (bit); sigalg: RSA-SHA256
v:NotBefore: Feb 3 17:19:36 2023 GMT; NotAfter: May 4
17:19:35 2023 GMT
... ssh-keygen -R github.com && curl 'https://api.github.com/meta' | jq -r '"github.com " + .ssh_keys[]' >> ~/.ssh/known_hostsAlso you should first check:
1. If you’re using GitHub’s rsa key at all
2. That you don’t have the other keys
In which case nuking all the keys is completely unnecessary, just remove the RSA one.
But how do you know it's the right key?
Funny enough, given that every person (and every bit of tooling) I know blindly approves SSH public key verification anyways, they will likely not even notice this switch.
Maybe on initial connect, but who TF ignores that "key has changed danger danger, high voltage" warning? You at least look around and ask a colleague?
If you don't pre-bake the known-hosts, then you'd allow each new ephemeral run to use whatever github tells you it's key is.
It did briefly break ours, as we pre-bake the known_hosts file into our CI image for convenience and security.
Convenience due to CI not having TTY's, so various tools would get stuck on prompt Y when connecting to github for the first time. (Which is every run, if you are ephemeral CI)
And security, as now everything broke due to github's key changing, which is the desired outcome actually.
We bumped the known_host key entry, merged and all is well again...
Given that the key is extremely long lived, this has unfortunate implications: If any of these servers is compromised, or decides to go randomly spewing memory content because of a bitflip, or screws up the nonce on a DSA/ECDSA operation, the key can be compromised. This is hard to exploit if you're a random person, but for a global adversary that collects internet traffic at scale (e.g. NSA), it's feasible and I would be surprised if they weren't exploiting such issues. They were collecting HTTPS handshakes for a reason.
You can't get two (well functioning) TPMs to have the same key. They come with their own, un-extractable* and unchangeable keys built in.
*TPMs claim this, it is probably not impossible to extract keys just incredibly difficult and requiring specialist knowledge.
All cloud vendors offer the same functionality, if you think about it, so it's not an obscure feature.
If anyone is using a Windows machine with TPM-based bitlocker encryption, you have followed the instructions at https://support.microsoft.com/en-us/windows/back-up-your-bit... I hope?
Users will need to actively remove the old rsa key in order to be safe. It's my first question, and a colleague suggested that they believe that the private key was not seen, however, I don't see that in the post - unless I'm missing it - and I really want this stated very clearly somewhere.
To be honest, I'd expect something like this to be mentioned in the announcement.
Yea that is my read on it as well. If that is true this is much more severe than the blog post suggests.
If they did something that expensive for security they would be bragging about it from their blog to earn any possible trust they could at a time when they could really use it.
If you are reading this Microsoft, make me look silly. Please.
I will even give you the enclave designs for free.
ssh-keygen -f ~/.ssh/known_hosts -R github.com
Note that github-keygen stores the ssh-ed25519 key since release 1.306 in June 2022. You can check by yourself:
cat ~/.ssh/known_hosts_github
curl --silent https://raw.githubusercontent.com/dolmen/github-keygen/release/github-keygen | perl
- An SSH key can be revoked on the server, but the client won't register it immediately, and all clients must manually verify the new key signature and update their local configuration. This doesn't happen with HTTPS; the client just works with the server's new cert, no user intervention is required.
- A username can be used in the HTTPS URI to tell both your client and the server what credential to use. SSH method requires the user to load the correct SSH key first.
- Most servers like GitHub allow more fine grained access control for HTTPS tokens than SSH keys.
- HTTPS access method works on all HTTP proxies, whereas SSH is often blocked.
- HTTPS can be faster than SSH.
- HTTPS allows a client and server to use specific ciphers to address regulatory and other requirements.
- HTTPS tokens are supported in a wider range of password managers/keychains than SSH keys.
- Most users don't password protect their SSH keys, but they often have a password manager with a master password that they can keep HTTPS token in.
- SSH key management is much more complex on the user end than HTTPS token management. The former requires ssh-agent, a key generator, and instructions for use, as well as specific filesystem permissions for keys.
- The user often doesn't understand the idea that they shouldn't share their private key. But everyone basically gets they shouldn't share their password.
- Virtually no one ever checks the fingerprint of ssh server keys. Often entire companies have configuration that disables host key checking, completely eliminating the security SSH is supposed to provide. With HTTPS the user is implicitly protected with no actions necessary.
- Nobody ever commits a private TLS key to a GitHub repo, but apparently they do with SSH private keys...
Are there password managers that actually seamlessly do that? GitHub's official docs (https://docs.github.com/en/authentication/keeping-your-accou...) for example recommends using the GitHub CLI to cache your authentication token for you and as far as I can tell the token is stored unencrypted (https://github.com/cli/cli/issues/1773), unlike SSH keys which should be encrypted behind a password. If you don't do this you have to enter your password every single time you use you Git, which is not good even if you have a password manager (it adds steps, and passwords aren't great to begin with because you send them in plain text to the server and it is why Apple etc are moving away from password-based authentication for websites).
> - Nobody ever commits a private TLS key to a GitHub repo, but apparently they do with SSH private keys...
I think the key issue here is this really shouldn't have happened. There is no reason why GitHub couldn't secure their SSH private key just like their TLS key.
---
Edit: Thinking more about it, I think the situation would have been much worse if GitHub leaked a TLS key instead.
With SSH, an attacker who stole GitHub's private SSH key actually *cannot* MITM the user. They can pretend to be GitHub to the user, but they cannot pretend to be the user to GitHub, because with SSH both sides have a private/public key known to the other side. In a normal usage scenario, it would actually take some social engineering to do a real attack, because a simple Git fetch from the attacker would fail when trying to fetch a private repo (since the attacker does not have the permission to fetch it from GitHub).
With HTTPS, under current setups of username/password or an auth token, the user has no real way to protect themselves from a MITM attack. The token isn't a private key as it's shared directly to the server. That means the attack could actually MITM the server/user connection and do much much more damage.
Oh how I wish this were true...
Also, SSHFP requires DNSSEC, which Github of course does not support. (amongst other shiny new technologies, such as IPv6... /r).
And even if GH would deploy DNSSEC, it'll still be opening you up to a host of other attack vectors that come with DNS based trust anchors.
That's bad advice. Everyone should remove the compromised key, not only people seeing the message. If you don't see the message and everything still works (while using RSA), you've been MitMed.
error pulling image configuration: download failed after attempts=6: x509: certificate is valid for *.githubassets.com, githubassets.com, not pkg-containers.githubusercontent.comLet's remove the words briefly exposed and public. Why is a private key in a Git repository? These days I don't see any reason for this.
Very opportune time for hackers to try this, since few are probably verifying the fingerprint.
How is this true? AFAIK, A MITM with the github priv key will be able to do an SSH downgrade attack...
> This week, we discovered that GitHub.com’s RSA SSH private key was briefly exposed in a public GitHub repository.
No further questions for this witness Your Honor
Something like that must have happened anyway, since it's highly unlikely a private key is just lying around as a plain text file on an engineers workstation to be accidentally included in a push.
> Could anyone with that key have cloned any GitHub repo, private or not?
If you try to push to GitHub but someone is MITMing you with the leaked key, they can say "I'm an empty repo" and the client will helpfully push all the commits to the attacker.
Until everyone updates their .ssh/known_hosts file, a MITM attack can still steal code in this way.
They seem to have been using the same key.
- Push security breach details down so fewer people read them. Nothing until line 11. Or line 13 if you count the TLDR.
- Litter your text with understatements. After publicly exposing keys, you replace them "out of an abundance of caution" to replace them.
And yet this is one of the best security breach posts I've seen in a while. When will companies start to respect developers?
Do the PR people have technical knowledge or does some tech person write
“I accidentally screwed up and put the key public so we have to replace” and the PR glosses over it and rewrites “Out of an abundance of caution…”?
Hyperlink with broken certificates. Fun.
I know I know. All you have to do is to just MITM someone's DNS or IP traffic or whatever and it's all over. Well, I will guarantee you that millions of bitcoin are sitting there waiting for you to take it. It's all yours. Go take it. You could be a millionaire tomorrow. And if you do it right, they'll let you keep some of it if you send most of it back.
I'd love to get a copy of the key, so I can setup a demo to encourage education.
> That said, if I'm honest I think almost no one checks the SSH host key. I would bet that for 9 out of 10 engineers / devops / security people, they will ignore the SSH Host Key and might even set StrictHostKeyChecking=no.
I believe you, and I agree this is common. I have been carrying the github host keys in my dot files for more than a decade. I do this somewhat ergonomically by having a Host config that adjusts the known_hosts path to permanent_known_hosts for things I care about.
Setting StrictHostKeyChecking=no is negligence, despite being common there really isn't two ways about that, it's a trivial vector with only a very moderate need for additional information to turn this into an RCE in most cases.
There are big problems with ssh's host keys, I understand and agree, but there are also big problems with passwords and we've made a dent in getting people to use password managers and 2FA. We should not wave this away just because it's a little harder, otherwise we're just relegating ssh to the same status as gpg.
Edit: where can I get good Mercurial hosting? F this.