Dan Demeter retweeted
Attending @DefCampRO next week? We have a treat for you! TLPBLACK researchers @craiu and @_xdanx will present on how we are "Fighting ransomware attacks with Arch Linux, AI, ***** and APT skills". Come and find out what's the other secret ingredient we use. Date: Thu 19 Nov · 12:05–12:50 · Europe/Bucharest at the majestic House of the People, aka' Ceausescu's Palace. def.camp/schedule/
Dan Demeter retweeted
The NSA mathematician who made the numbers protecting your passkeys died in 2023 without ever remembering how he made them. Cryptographers chased the answer for years and put up a $36,864 bounty. Claude figured it out in under 30 minutes, and the answer was the clock on his computer.
His name was Jerry Solinas. Back in 1997 he generated the starting numbers for P-256 and a few sibling curves, the math underneath your passkeys and a big chunk of the secure websites you visit every day. NIST published them in 2000 with no explanation of where they came from, and for a long time people mostly shrugged.
Then 2013 happened. The Snowden files showed the NSA had slipped a backdoor into a different crypto standard, and suddenly every unexplained number the agency had ever handed out looked suspicious. These were the most widely used ones of all.
Jerry's own story was that he'd hashed a joke sentence, something along the lines of "Give Bob and Jerry a raise," and then lost it when his work computer got swapped out. He tried to recreate it himself and never could. He even joked that any backup was probably sitting in the same warehouse as the Ark of the Covenant.
After he died, Steve went digging. He tracked down former NSA people and old cryptographers, and four of them independently told him the same joke-phrase story. Filippo Valsorda turned that into a bounty in 2023, and for three years people threw computing power at every version of that sentence they could dream up. A few folks at Anthropic took a crack at it too, Steve among them. He's now one of the people credited on the solve.
None of it matched.
So someone at Anthropic tried something different. They asked an internal Claude model to figure out where the seeds came from and purposely left out the joke-sentence clue.
Claude's first move was to think like a programmer in 1997. If you needed random numbers back then, you probably grabbed the basic generator that came with the C library on your Unix machine, and those old generators are simple enough that they leave fingerprints in their output. Sure enough, in four of the seeds the top bit of every 32-bit chunk was a zero. The odds of that happening by accident are about 1 in 5,810, and it's the calling card of a generator called lrand48.
And this is where it gets fun. lrand48 runs in reverse just as easily as forward, so Claude rewound each seed back to the moment the generator was switched on. Every prime-curve seed rewound to a date and time from 1997 or 1998. Whoever ran the program had seeded it with the computer's clock.
P-256's run started at 1:52:34 AM UTC on August 16, 1997. If that machine was at NSA headquarters in Maryland, somebody hit enter at 9:52 on a Friday night, and the seed that ended up in the standard was the 1,842nd one the program spat out.
Jerry's joke sentence was never in there. He misremembered, four people faithfully repeated what he told them, and some very good cryptographers spent years trying to crack a hash of a sentence that nobody ever typed. Leaving the clue out is probably what let Claude get there in half an hour.
I love the irony here. Using the clock to seed a weak random generator is one of the first things they tell you never to do in a security class. That exact laziness is why anyone can now rerun the NSA's 1997 program on a laptop and get the identical numbers, which makes it really hard to argue anything sneaky was hidden inside them.
There's still one thing left to argue about. In the P-256 run, number 876 also would have made a perfectly good curve, and someone skipped it and kept going to 1,842. Anthropic says so few got skipped that it couldn't have been a hunt for a weak curve, but nobody knows why it happened.
Claude's own notes from that moment are pretty funny. When the numbers matched, it actually stopped and wondered whether it was inside a test, since a 25-year-old mystery falling apart that fast seemed too good to be true. Then it reran the logic and talked itself into it, because ten curves each regenerating perfectly from a single timestamp doesn't happen by luck.
The $36,864 is going to the nonprofit that runs Let's Encrypt, the group that gives free security certificates to millions of websites, a lot of them built on these same curves.
CLAUDE FOUND THE SEEDS
anthropic.com/research/ecdsa…
Dan Demeter retweeted
GTIG has just released our latest quarterly AI threat tracker, a collection of case studies based on our unique visibility into this rapidly evolving threat. Criminals and state actors are making strong headway in speed, scale, and sophistication. 1/2 cloud.google.com/blog/topics…
If you're into AI honeypots, this is a funny story
"Ollama, vLLM, all the homegrown agent stuff people are exposing without thinking too hard about it - are a genuinely new kind of thing to attack"
Our latest articole by @marco_preuss is up!
tlpblack.net/blog/20260804-h…
Dan Demeter retweeted
"Ollama, vLLM, all the homegrown agent stuff people are exposing without thinking too hard about it - are a genuinely new kind of thing to attack"
Our latest articole by @marco_preuss is up!
tlpblack.net/blog/20260804-h…
Quite an interesting case we've analyzed.
We expect this malicious threat actor to continue its activity in the future.
This is our analysis based on the ransomware incident reported by Romanian's Land registry agency, pivoting on the reported IoCs.
Our latest blog post "A Deep Dive into ByteToBreach's ByteToCrypt ransomware" is here!
There are multiple quirky aspects of this bytetocrypt and the known actor's TTPs, suggesting that this ransomware was developed by a capable software developer, but less experienced at ransomware.
- the encryption implementation is fragile and error-prone but not cryptographically breakable
- this code is not an ESXi-aware platform but has been deployed to ESXi environments
- the ransomware is not runtime interactive for its operator
- following encryption and ransom tasks, it autoexecutes a set of anti-forensics capabilities
This immaturity invites extra attention to bytetocrypt's encryption capabilities. But when it comes to its implementation, all-in-all the technology presented is brittle but effective. Our assessment of the ransomware is based on several factors:
- it ignores randomness, cryptographic, I/O (deletion and rename), and anti-forensics failures
- mistakes read errors for end-of-file
- truncates paths to fixed buffers
- loses metadata
- can announce success after partial or failed work
Continue reading our analysis: tlpblack.net/blog/20260728-a…
4
848
Dan Demeter retweeted
This week's problem is out on all platforms, presented by @ThinkstCanary
open.spotify.com/episode/5Uk…
Dan Demeter retweeted
Our team is back from #CybersecEurope - which took place this week in the beautiful city of Brussels. This is Belgium’s largest IT and cybersecurity fair, focused on IT security, data management, cloud, AI, and digital resilience under the theme “Stronger Together”. It was a solid, professional industry gathering with a focus on collaboration, practical strategies, and Europe’s push for cyber sovereignty amid rising threats. Thanks to everyone who came to talk to us and was great seeing you there!
Dan Demeter retweeted
The NATO Locked Shields 2026 exercise proved (again and again) that the real challenge is not only technical, but also people management. Coordinating a team of more than 300 professionals is not an easy task. However, we believe this is the beauty of this exercise: it creates a safe space to fail, learn, and grow - so that real-world readiness is earned, not assumed. Secondly, jumping into the live-fire arena, where seconds matter and mission-critical systems (such as SCADA systems controlling water, electricity, or even voting systems!) have to be online with as few disruptions as possible is an environment that few events or exercises manage to simulate at this large scale.
We would like to thank all the people who worked with us late at night, patching, testing, reverting VM snapshots, reversing malware and writing Yara rules. You know who you are :)
Lastly, we would like to thank all organizers, both from Romania and the NATO Cooperative Cyber Defence Centre of Excellence (@ccdcoe) for organizing this event and we are looking forward to participating in next year's Locked Shields!
Read the full story on our blog: tlpblack.net/blog/20260427-l…
(Photo taken where all the magic happens. Somehow a reminder that the sky is the limit!)
I wasn't born then, but I remember my parents' stories about Iodine pills
минуло сорок років.
але я звертаюсь до вас.
до тих, хто тоді, не вагаючись, крокував у це пекло. до воїнів, які не мали перед собою видимого ворога. не було прапорів супротивника, не було чутно вибухів, лише дивний металевий присмак на губах і наказ, який треба було виконати.
ви гасили пожежу, що випалювала клітини вашого життя. хтось із вас скидав графіт із даху, рахуючи секунди, що коштували років майбутнього здоров’я. хтось копав тунелі під реактором, відчуваючи, як піт застигає на чолі разом із радіоактивним пилом.
ви не думали про пільги чи посвідчення, які потім доводилося вибивати в кабінетах байдужих чиновників. ви просто робили свою роботу, бо знали: за вашими спинами - діти, жінки та світ.
ваші серця виявилися міцнішими за бетон і графіт. ви стали тим живим саркофагом, який закрив собою людство від розщепленого атома.
багатьох із вас уже немає. ви пішли тихо, один за одним, згорівши від того невидимого вогню, який прийняли на себе в далекому 86-му.
я схиляю голову перед кожним пам’ятником на кладовищі та перед кожним із вас, хто сьогодні ще тримається.
сьогодні світ часто забуває про вашу жертву, занурюючись у дріб’язкові суперечки.
але я пам’ятаю. і я хочу, щоб ви знали: ви і є атланти, які тоді втримали небо над нами.
дякую!
109
Dan Demeter retweeted
Finally, this research is dedicated to the memory of our late friend Sergey Mineev (@Skvern0), APT hunter extraordinaire, who pioneered the techniques we used to get here. He would’ve loved this research.
Happy hunting!
Dan Demeter retweeted
Research dedicated to mr Sergey Mineev @Skvern0, who pioneered many of the techniques that enabled this discovery.
Dan Demeter retweeted
Pretty much everybody who looked at it before had missed it - cool research here by our friends @juanandres_gs and @vkamluk - wired.com/story/fast16-malwa…
Dan Demeter retweeted
The name ‘fast16’ is referenced in the infamous ShadowBrokers’ leak of NSA’s ‘Territorial Dispute’ components.
fast16.sys selectively targets high-precision calculation software, patching code in memory to tamper with results. By combining this payload with self-propagation mechanisms, the attackers aim to produce equivalent inaccurate calculations across an entire facility.
sentinelone.com/labs/fast16-…
wired.com/story/fast16-malwa…
Best paper I've read so far this month:
All elementary functions (sin, cos, tan, exp, log, powers, roots, hyperbolic functions, π, e, and even basic arithmetic) can be generated from just one binary operator:
eml(x, y) = exp(x) − ln(y)
…plus the constant 1.
This Polish theoretical physicist just proved you can recreate all math functions from JUST one operation.
E(a, b) = e^a - ln(b)
Every single operation: +, -, x, / , trig, log, as you can see below.
Extremely mathematically elegant.
Dan Demeter retweeted
Bias in CTI attribution is harder to shake than you think.
The DPRK/cybercrime boundary has always been blurry, but today I got a personal reminder of just how biased I still am.
Here's what happened:
1⃣ First look: I encountered a compiled AppleScript malware (Zoom Audio SDK Update.scpt, MD5: 743f60f02d352201e2df36805f1ec00d). I suspected a link to `APT38` due to their known use of similar social engineering lures, including ClickFix techniques.
2⃣ Deeper dive: I checked the 2nd and 3rd stage payloads. I told myself they'd just changed their tradecraft. Still convinced it was DPRK.
3⃣ Final payload: The last-stage AppleScript was focused on stealing a wide range of sensitive data from the victim. Something felt off. I ran it through my IOC search bot. Result: `SHub Infostealer`, known crimeware, nothing to do with DPRK. DataDog published a solid write-up on it: 🔗 securitylabs.datadoghq.com/a…
If I had stopped at stage 1 or 2, I would have published a completely wrong attribution. Attribution requires full context. Our prior knowledge is a lens that helps us work faster, but it's also a blindfold. The more deeply you specialize in one actor, the more you'll see that actor everywhere. Stay humble. Follow the evidence to the end.
Dan Demeter retweeted
🚨 Following the #CPUID compromise, we've been digging into the payload delivered by the threat actors.
The payload is a trojanized DLL abusing DLL-sideloading to deploy the next stage.
It reuses the same C2 config and staging server as the loader reported last month by @Malwarebytes, which was pushing a trojanized FileZilla client via the same DLL-sideloading trick.
Using our code-similarity engine, we identified the next stage as #STXRAT — a new RAT with infostealer capabilities, documented by @esthreat 2 days ago.
🧬 #IOCs:
49685018878b9a65ced16730a1842281175476ee5c475f608cadf1cdcc2d9524 (Trojanized DLL)
welcome[.]supp0v3[.]com (C2 staging server)





























