Das Internet ist in höchster Alarmbereitschaft, nachdem acht kompromittierte TLS-Zertifikate entdeckt wurden
https://arstechnica.com/security/2026/10/the-internet-goes-on-high-alert-after-discovery-of-8-compromised-tls-certs/
Das Internet ist in höchster Alarmbereitschaft, nachdem acht kompromittierte TLS-Zertifikate entdeckt wurden
https://arstechnica.com/security/2026/10/the-internet-goes-on-high-alert-after-discovery-of-8-compromised-tls-certs/
44 Kommentare
What certs and by which issuer?
Wish this was written so everyone (non technical folks) could understand what’s actually happened.
> The private keys of eight eMudhra intermediate CA certificates have been compromised. They chain to roots trusted by the Mozilla, Microsoft, Apple and Chrome root programs and by many downstream trust stores.
Ho-lee-jeezus.
Really shameful by eMudhra – they’ve squandered their community obligation with such reckless process allowing this to happen. Yikes.
So is using ProtonVPN gonna be riskier now? At least for some time?
For non tech people, TLS is an encryption method (or collection of methods). Modern encryption basically works by converting the text or data into a number and then multiplying that number by 2 (or more) very large prime numbers.
To decrypt the message, you divide by the prime numbers. Two large numbers multiplied give you a much larger number, and so checking which prime numbers make up the big number is a very time consuming task unless you know what the answer is – that is, you have the decryption key.
There are other reversible math operations that are easy to perform in one direction but difficult to reverse unless the key is known. The article uses something slightly different but im not super familiar with the details.
But the secret key numbers that were supposed to be known only to the secure party were (4, 56, 66, 115, 155, 156, 230, 237).
These are… very very small for a secret key. The article states that the encryption method allows you to have 116 digit numbers in your secret key.
Its likely that it was test code that sneaked its way into live environments by accident.
Looks like a white-hat caught it and they’ve since revoked the key. Your device is still vulnerable to this if you havent updated the applicable devices, especially in places with public wifi, or other devices that connect your device to the internet
The internet should see me Event Viewer. Closed and done with that shit
hopefully none of the election infrastructure needed https
Only the beginning…
Well… that’s not good. Yikes. I am still coming across devices which are vulnerable to Heartbleed and will never be patched.
This also brings me back to the days where malware on home PCs would exploit home routers with default passwords or vulnerable firmware and modify the DNS servers to ones controller by an attacker. Doing that allowed the attacker to point requests for certain websites to their own servers. Or, just return false data like they would for Windows Update and any Anti-virus vendor to ensure a PC remains vulnerable. They could combine a similar exploit with this vulnerability to get after IoT hardware and other equipment which will probably never see a patch.
On the flip side, this same exploit could be a boon for jailbreaking hardware.
Stuff like this reminds me of prior conversations on reddit where people were so insistent on tls being this infallible security layer. “Oh don’t worry about using the wifi on a hotel network, as long as you’re using tls https blah blah blah.” Heavily upvoted. Then comments about this kind of stuff being possible, massively downvoted. Goes to show the crowd doesn’t always know especially on bigger subs.
We definitely need something better than certificates to validate trust in 2026. Certificates often require manual intervention to update, which makes situations like these unbearable. The encryption is fine, but the distribution is crazy manual and cumbersome for 2026.
Non-tech guy here, semi-freaking out. So if I’m understanding correctly, these security keys act as ID cards, and in the wrong hands can be used to disguise one website as another? Like, if I typed in „www.Google.com,“ the website that I’d be taken to could be a lie disguised as Google?
And how old are these keys? Is this a „many things in the past few months may have been compromised“ scenario, or „fuck shit fuck *everything since the Internet started* is compromised“ scenario?
Can someone explain like im 5
For some context: The company responsible is **eMudhra**, based in India. These 8 dangerously weak security keys were generated in April 2025 and went undetected for nearly **18 months**, until an independent researcher exposed them on October 17. The astonishing part is that their weakness was trivially detectable, raising serious questions about the company’s quality control and security audits. **The actual danger** is that someone could forge trusted website certificates and potentially intercept communications, provided they could also position themselves between you and the website. **It does NOT mean everyone’s passwords, banking information or Internet traffic have been compromised.** No malicious exploitation has been confirmed. Major browsers have already blocked the affected certificates, although older software and embedded devices may remain vulnerable until updated. eMudhra has revoked the certificates and is investigating its key-generation procedures; the root cause and accountability remain unresolved.
Kudos to the commenter who used the passport analogy — an excellent way to explain the problem without drowning everyone in technical jargon.
So on a scale of „mildly concerning“ to „Bring me my brown pants“, where would this fall?
The company I worked for in ~2018 got caught in a large scale distrust and revocation event. It might have been Symantec or GoDaddy, I can’t remember. We were notified overnight we had 12 hours and I saw the email that morning; we replaced our entire certificate stack in less than half a day, all hands to the pumps. I think that was the most urgent P1 I ever was involved in.
Lol of course it’s India.
benafflecksmoking.bmp
Oh wow, this is a great way to demonize VPNs so people stop using them.
My new company is called:
analog everything.
“We don’t want your data. Hell, I don’t even wanna know your name”.
We’re gonna sell washing machines and dryers with knobs and buttons, and parts you can install yourself if so be it old school clamshell cell phones that only make phone calls and T9 text video game systems you have to blow into the cartridge to make work.
thank you india.
hey these attack vectors are sophisticated and I’m pretty sure a major proliferation of them is by using spoofed Gmail log in linked account generators for compromised 3rd party sites.
I’m a construction estimator and I received an email that was, at first glance, a super legit appearing ITB from an unfamiliar GC (not uncommon). The link provided in the email sent me to an iSqFt construction document file hosting / viewer platform with a seemingly legit gmail linked account creation portal. My at the time chrome browser version showed that the certs were trusted but the URL was something odd like “elegalumin” dot com and nothing like ive ever seen with similar situations. Ultimately I didn’t trust it, backed out and reported to gmail as suspected phishing attempt.
One week later my boss tells me his “gmail account has been compromised.” Didn’t elaborate any further but this now has me wondering
This site checks to see if your device is still accepting the compromised certs: https://emudhra.we-have-a-problem.com/
Does this affect apps on my phone? Say if I have proton vpn app on my phone could it have been vulnerable to an automatic update that installed a dummy application?
Sorry for the likely dumb question but I’m a very simple layman when it comes to this and not entirely understanding the issue at hand
Just a reminder that if you weren’t expecting an email from a site about your account, it’s better to open the site separately and directly before logging in. These days anyone can spoof an email and make it look real enough..
This is fucking hilarious on so many levels:
>The compromised eMudhra keys were defective at a level that’s challenging the imagination of many of the engineers studying the incident. A normal size for a NIST P-384 elliptic curve cryptography private key contains roughly 384 bits of entropy, a size allowing almost 2384, or a number with 116 digits, different possibilities. The private key sizes of the compromised keys ranged from 3 to 8 bits. Astonishingly, the private keys were 4, 56, 66, 115, 155, 156, 230, 237.
The private key, which is normally a huge number no one would be able to keep in their head, was 4. **4.**
Imagine only being able to get to any location without GPS. Then your GPS get hijacked and instead of going to your bank, your GPS sends you to a fake bank run by thieves.. or a fake grocery store that has insanely cheap prices for everything you’re looking for..
except all of it looks real, your apps tell you its legit.
The Ars Technica article confirms that the security alert stems from eight compromised intermediate TLS certificates belonging to India-based certificate authority eMudhra Limited.
The reporting party proved they had control of the keys by digitally signing an informal advisory using the private key of one of the certificates.
The core details from the article align perfectly with the broader systemic impact:
The Core Flaw: The private keys were astonishingly weak, ranging from just 3 to 8 bits in size (with specific private key values revealed as 4, 56, 66, 115, 155, 156, 230, and 237). A normal secure key requires 384 bits of entropy, leading experts to call the vulnerability absolutely „bonkers“.
The Heartbleed Connection: Cryptography engineer Filippo Valsorda noted that while browsers blocked the certificates rapidly, updating the rest of the web (like IoT devices, Docker containers, and forgotten Linux machines) will be a massive cleanup job with a „very, very long tail“—potentially making it a larger logistics hurdle than Heartbleed.
Another day in India.
So glad I’m OOO for the next 3 days.
One of the passwords they used was „4“, hahaha.
It would have been „5“ but the admin was off sick that week from B-tech and didn’t realize you could count the thumb too.
***THE*** Internet?!
Convenient timing for the mid terms tho
This is company killer level of fuck up. Every corporate security team and open source project are furiously applying mitigate today. This company would not be trusted as CA
So who decided these people get the authority to issue certificates? Clearly they are incompetent
“My guess is that it’s example code that accidentally leaked to production, but hard to know. But it’s bonkers that no one noticed.”
You’re absolutely right! The secret should be more than 4 characters long! I was just using a simple example if this for a Certificate Authority you should definitely not run this code in production!
Almost definitely a silly question as I haven’t bothered to look it up but I’ll ask in case anyone else is curious; is this at all related to the critical php vulnerabilities that were just recently reported? I wouldn’t think they are but I know very little about certain authorities, maybe somehow its related
Alright. To heck with it, ill be the one to ask.
Can someone simplify this so someone who can only basically run the surface level stuff online, and understand that same surface level stuff, can understand this?
*The La li lu le lo!*
Oh maybe this is how someone stole my digital identity at every single level lol FML
Huh. Apparently on Android you can just go into settings and disable all eMudhra root certs.
Pretty sure I can do the same on windows when I have time.
Linux is the only one that might be problematic as they’d be replaced on updates, I need to see if there’s methods to mask them for exclusion 🤷.
Firefox, Ubuntu (24.04) and Debian 13 have patched this, no clue yet on Rocky Linux, Alma or RedHat!
AI code being pushed to prod without review, perhaps?