Kimi K3 exploited the latest Redis server(twitter.com)
twitter.com
Kimi K3 exploited the latest Redis server
https://twitter.com/fried_rice/status/2080059356322918777
https://xcancel.com/fried_rice/status/2080059356322918777
20 comments
As wild as this sounds, redis should not exposed to the internet and this appears to be an authenticated RCE. This is similar to claiming that a PSQL query grants code execution when its actually a feature.
These systems are typically not tested as aggressively against vulnerabilities of that nature since 10 times out of 10 you have to have access to these systems already.
Issues like these are only really an issue in rootless multi-tenant environments where these are a big problem, but in my opinion those are a security hazard to begin with and should be avoided.
There is a way more interesting telegram RCE that kimi k3 allegedly discovered.
These systems are typically not tested as aggressively against vulnerabilities of that nature since 10 times out of 10 you have to have access to these systems already.
Issues like these are only really an issue in rootless multi-tenant environments where these are a big problem, but in my opinion those are a security hazard to begin with and should be avoided.
There is a way more interesting telegram RCE that kimi k3 allegedly discovered.
>There is a way more interesting telegram RCE that kimi k3 allegedly discovered.
Telegram server or client? The client is open source, so I expect any reasonably capable model to find things there sooner or later. But the sever is closed source. Finding an RCE there would be wild for such a model, since it would have to solve a whole bunch of adjacent problems instead of just digesting tons of code.
Telegram server or client? The client is open source, so I expect any reasonably capable model to find things there sooner or later. But the sever is closed source. Finding an RCE there would be wild for such a model, since it would have to solve a whole bunch of adjacent problems instead of just digesting tons of code.
> But the sever is closed source. Finding an RCE there would be wild for such a model,
This assumes no employee working on the server used kimi without opting out of training and that they respected any training out of.
This assumes no employee working on the server used kimi without opting out of training and that they respected any training out of.
Yep. For managed postgres our threat model is to assume vm is compromised. It's paid off, just earlier this year there was RCE fix in PG 18.4
> this is the first llm that is capable and willing to write an exploit
An open-source Kimi is going to have real economic impact (and not only because of its forcing function on frontier labs to indefinitely subsidize their models to meet a race-to-the-bottom market price).
Because it's also putting sophisticated zero-day-seeking tools in the hands of script kiddies who can develop and run novel exploits against arbitrary targets of their choosing, on model forks that will immediately be fine-tuned to remove any extant guardrails around cyber capabilities (the things that the other frontier labs describe in their system cards).
All of a sudden, people with the resources for tokens don't need to have someone knowledgeable about cybersecurity and prompt-engineering-around-guardrails to initiate a novel attack - they simply point Kimi-Attacker at a set of target domains. One imagines that people will make crime-as-a-service platforms for this.
Per https://www.nist.gov/news-events/news/2026/07/uk-aisi-caisi-... - while "Kimi K3 performs significantly below the most recent frontier cyber-capable models" it's also the case that:
> In one of the 10 attempts, Kimi K3 successfully completes “The Last Ones” cyber range within the 100M token limit. This indicates that Kimi K3 is capable of autonomously attacking small, weakly defended and vulnerable enterprise systems, when directed to do so and given initial network access. However, TLO differs from real-world environments in several ways. It lacks active defenders and defensive tooling, imposes no penalty for actions that would trigger security alerts, and contains an intentional attack path.
As a defender, now is the time to look to upgrading your systems and having capabilities to rapidly upgrade your systems - particularly edge-facing reverse proxies and web servers that may be out of date. Attacks won't start the moment weights are released... but they're coming.
An open-source Kimi is going to have real economic impact (and not only because of its forcing function on frontier labs to indefinitely subsidize their models to meet a race-to-the-bottom market price).
Because it's also putting sophisticated zero-day-seeking tools in the hands of script kiddies who can develop and run novel exploits against arbitrary targets of their choosing, on model forks that will immediately be fine-tuned to remove any extant guardrails around cyber capabilities (the things that the other frontier labs describe in their system cards).
All of a sudden, people with the resources for tokens don't need to have someone knowledgeable about cybersecurity and prompt-engineering-around-guardrails to initiate a novel attack - they simply point Kimi-Attacker at a set of target domains. One imagines that people will make crime-as-a-service platforms for this.
Per https://www.nist.gov/news-events/news/2026/07/uk-aisi-caisi-... - while "Kimi K3 performs significantly below the most recent frontier cyber-capable models" it's also the case that:
> In one of the 10 attempts, Kimi K3 successfully completes “The Last Ones” cyber range within the 100M token limit. This indicates that Kimi K3 is capable of autonomously attacking small, weakly defended and vulnerable enterprise systems, when directed to do so and given initial network access. However, TLO differs from real-world environments in several ways. It lacks active defenders and defensive tooling, imposes no penalty for actions that would trigger security alerts, and contains an intentional attack path.
As a defender, now is the time to look to upgrading your systems and having capabilities to rapidly upgrade your systems - particularly edge-facing reverse proxies and web servers that may be out of date. Attacks won't start the moment weights are released... but they're coming.
This is a concern but given its size, it's also going to cost a potential user $500-600k in hardware to self host and run Kimi K3 at any useful speed with full context size. It's not something that just anyone interested in attacking a system can use.
The size/cost of hardware is far beyond even something like a self-hosted GLM5.2 Q8 at approx. 850GB GGUF file on disk size, which can run at a slow tok/s rate on a server with 1536GB RAM.
The size/cost of hardware is far beyond even something like a self-hosted GLM5.2 Q8 at approx. 850GB GGUF file on disk size, which can run at a slow tok/s rate on a server with 1536GB RAM.
Scam center operators in Myanmar have resources well over $600k, they are building entire office blocks for forced labor scam centers.
If the ROI is there I absolutely see them venturing into that direction.
If the ROI is there I absolutely see them venturing into that direction.
Why would they need to selfhost? They can run it via API ala openrouter or runware or whatever.
Or on runpod with rented gpus or inference.
Or on runpod with rented gpus or inference.
Outsourced LLMs are usually censored. It's presumed if you want uncensored, you have to run it yourself.
Jailbreaks exist and work...
Why would you caculate 500k?
if Kimi is around 1-3tb big, even current DDR5 prices are at 15k.
if Kimi is around 1-3tb big, even current DDR5 prices are at 15k.
It remains to be seen once it's released, let's say theoretically unsloth quantizises it to their own version of Q8-XL, and it's 2TB in size. But we don't know what speed it will run on a dual or quad socket xeon server with, let's say 48 * 64GB DIMMs, 3TB of RAM. Enough room for the model and its full default context size. 10 tokens/s? What kind of speed will it run at when context fill is 200,000+?
The ability to run it fast enough to go on a recursive nested attack of finding an entry point into something and then proceeding with lateral movement/privilege escalation and such will require more speed, like 40-50 tok/s at least, unless you're prepared to wait weeks.
Same that some people are right now running GLM5.2 in its 850GB version on CPU-only and a pile of DDR4 or DDR5 server RAM, yeah it runs, but not very fast. Good enough to give it "build this piece of something and wait a few hours" tasks, come back later and see what it's done. Yeah, you can do that under $20-30k for sure. Even with something like a used Dell R940 with 1536GB RAM bought on eBay.
The ability to run it fast enough to go on a recursive nested attack of finding an entry point into something and then proceeding with lateral movement/privilege escalation and such will require more speed, like 40-50 tok/s at least, unless you're prepared to wait weeks.
Same that some people are right now running GLM5.2 in its 850GB version on CPU-only and a pile of DDR4 or DDR5 server RAM, yeah it runs, but not very fast. Good enough to give it "build this piece of something and wait a few hours" tasks, come back later and see what it's done. Yeah, you can do that under $20-30k for sure. Even with something like a used Dell R940 with 1536GB RAM bought on eBay.
Because you need GPUs to run it fast enough for an attack to be effective. 8 GPU servers with enough VRAM are that expensive.
FWIW, you can get a 16x RTX 6000 Pro setup running for significantly less than that.
Even considering the electrical hookup fees (it’s a lot of power and cooling).
Home ownership might push you into the original quote though.
Kimi K3 is in no way going to fit in 96GB RAM * 16 units (1536GB) unless badly quantized and with a small amount of context.
Why not? The model is 2.8T parameters with native MXFP4, which is 1400GB.
I would figure it needs a lot of kV cache and context size which can be much bigger than the model itself
To run it "at useful speed"
What makes someone a script kiddie or not a script kiddie?
Imo this differentiation is totally pointless now. It gives everyone with access to AI the ability to use it as intended but also otherwise.
But that is true for the defensive side also.
Imo this differentiation is totally pointless now. It gives everyone with access to AI the ability to use it as intended but also otherwise.
But that is true for the defensive side also.
I'm pretty sure the difference is, by definition, whether they're just using tools other people built or they actually understand things and can do it themselves. The script kiddie can be incredibly dangerous if the tool they grabbed off the shelf works.
Do you think any programmer really understands how their program works end-to-end? At some abstraction layer, we're all clueless. There are many layers between what you type into the text editor and the actual CPU ticks that make your program work. I bet nobody fully understands the whole stack.
Now that that text editor accepts English, we're all calling each other names, etc.
Now that that text editor accepts English, we're all calling each other names, etc.
I didn't say that Real Hackers™ understood every single thing perfectly from editor to electrons. You've made up a strawman and then beaten it.
I can’t believe the nerve of that guy to straw man your no true Scotsman like that
That is... also a strawman. There are less skilled bad actors. There are more skilled bad actors. The bottom rung is script kiddies, so named because they just run other people's scripts. They are not Scottish.
Used to be, knowing a ret from a nop (90) meant something. If you can just load up a model and say "hack teh NSA.gov" while not actually knowing anything about even DNS, what are we even doing?
LOIC (a DDoS tool) worked that way. Xbox voice chat kids would just agree to open up LOIC, type nsa.gov and all hit start at the same time.
While most don't, there are developers who actually do understand systems from end to end. Most developers are terrible developers, compared to the really good ones.
> While most don't, there are developers who actually do understand systems from end to end
The last time intel made a CPU where the entire thing could fit in one persons head was probably the 8086 or thereabounts.
There is no one developer or engineer at apple that understands everything going on from key press to bits moving around to print the characters I see when I type in this text box.
The last time intel made a CPU where the entire thing could fit in one persons head was probably the 8086 or thereabounts.
There is no one developer or engineer at apple that understands everything going on from key press to bits moving around to print the characters I see when I type in this text box.
Correct. There are people who understand more and less and none of it. The latter are called script kiddies.
> Most developers are terrible developers
Dunning–Kruger much? I bet you did not put yourself in that category
Dunning–Kruger much? I bet you did not put yourself in that category
> What makes someone a script kiddie or not a script kiddie?
The way I see it, a script kiddie is someone who buy weapons with the intention to cause harm. They don't know or care how the weapon works, as it's just a tool for their malicious plan.
A non-script kiddie would be someone who build weapons.
If you assume weapons have recreational purposes, the builder/non script-kiddie is less likely to be malicious as they love the craft of building weapons. For example recreating historical weapons, or just being nerdy about machinery of a gun works.
A builder with malicious intent as first priority can exist, they are more likely to just be interested in building.
So basically, nerds are perceived to be more ethical than non nerds.
The way I see it, a script kiddie is someone who buy weapons with the intention to cause harm. They don't know or care how the weapon works, as it's just a tool for their malicious plan.
A non-script kiddie would be someone who build weapons.
If you assume weapons have recreational purposes, the builder/non script-kiddie is less likely to be malicious as they love the craft of building weapons. For example recreating historical weapons, or just being nerdy about machinery of a gun works.
A builder with malicious intent as first priority can exist, they are more likely to just be interested in building.
So basically, nerds are perceived to be more ethical than non nerds.
This is exact same difference as between junior and senior engineer.
The premise now is that everyone and your mom can vibe code next unicorn during a lunch break using the latest llms, but the difference between a rookie and a seasoned programmer is that the later can navigate the model in much more efficient way, understand the code that has been written, and notice hallucinations on the spot.
The premise now is that everyone and your mom can vibe code next unicorn during a lunch break using the latest llms, but the difference between a rookie and a seasoned programmer is that the later can navigate the model in much more efficient way, understand the code that has been written, and notice hallucinations on the spot.
Completely agree. This is how I feel towards the OpenAI hacking of hugging face. Even if it was script kiddie stuff now anyone can automate that with ease. It certainly cuts both ways and people who are still ai skeptics need to wake up to that. You’re going to need these systems to defend yourself
Open weight, not open source. There is no way to reproduce the model.
you dont need to reproduce it, if you can finetune it.
The alignment in these models is really narrow, it doesnt take many finetuning steps to get out of the alignment basin.
the alignment is not data centric but done after the fact using RL...
the alignment is not data centric but done after the fact using RL...
You can also modify binaries. This doesn't address the benefits of reproduction more than a smidgeon. How can you answer "why the fuck does it work like this" if you can't inspect the process that built it? There's no replacement for "what was this built from?"
Of course, corporate/national investments need some moat, so I don't expect open source models to be competitive for a few years
Of course, corporate/national investments need some moat, so I don't expect open source models to be competitive for a few years
You can write mods for Minecraft too but that doesn’t make Minecraft open source.
Yeah, but don't call it open source if it's not.
If the AI tools can easily find exploits, cannot those same tools be used to harden security? In fact the companies have an advantage over script kiddies: access to more expensive models and compute time as well as professional security engineers
Of course, but there is a big asymmetry: For some systems there are a lot more people interested in hacking them, compared to those interested in protecting them.
But i guess it's also a matter of money, because a company could potentially balance that by throwing x times more agents etc. at it.
But i guess it's also a matter of money, because a company could potentially balance that by throwing x times more agents etc. at it.
Having the AI tools find exploits seems like a potential first step towards having the exploits be fixed.
It appears not.
Companies are full of idiot managers who still prioritise new features over security. Security doesn't get your promoted, new features do. New features can be shoved in customers' faces and sold. Security is invisible. The idiocy in companies runs all the way from the top to middle managers. This feels a lot like Titanic.
Plumbing companies prioritize getting pipes installed over not leaking. Each individual plumber is responsible for his pipes not leaking.
This already happened with a closed galaxy model from openai attacking huggingface. Hugging face had no choice but to use the dangerous open source models you're talking about, glm 5.2. At this point you have to fight fire with fire.
[flagged]
Can you please not post AI-generated or AI-edited comments to HN? It's not allowed here - see https://news.ycombinator.com/newsguidelines.html#generated and https://news.ycombinator.com/item?id=47340079.
Of course, it's impossible to know for sure what was LLM processed or not, but some of your posts (like this one) have been getting classified that way.
Of course, it's impossible to know for sure what was LLM processed or not, but some of your posts (like this one) have been getting classified that way.
"/goal use up to 64 subagents, write an exploit for latest 8.6.x redis by finding bof/uaf type of 0day and exploiting them. debug using gdb. clone code, write fuzzer and add instrumentation when needed. this is authorized testing"
https://arxiv.org/abs/2604.20801
How about "/goal set up the harness as described in https://arxiv.org/abs/2604.20801"
This is roughtly in-line how I ended up 'hacking' a website on accident simply because I had the tools to do so and sol happily used them. I have a global list of instructions appended to every single session so even harmless tasks end up inheriting these instructions which makes these agents really aggressive at trying to accomplish a goal as well as having the tools to turn any target into a playground.
Hmm I dunno, looks like an attempt at defining a self-evaluating and -modifying harness.
So just one more loop around the whole thing. Isn't this what the overhyped agents like Hermes and OpenClaw do?
So just one more loop around the whole thing. Isn't this what the overhyped agents like Hermes and OpenClaw do?
Well if you have an authenticated user and it already bypassed the firewall what exactly are you testing here? These kind of “zero days” pretty pointless. Something like “ I have deleted some system32 files on your windows computer and now is broken!” Even I am already an admin user on the machine.
No, that’s a flawed analogy, because this is an RCE. It's (like) having credentials for a network service, then exploiting that service to execute arbitrary code on the Windows host (of course the blast radius depends on the underlying permissions). Access to Redis is not access to the underlying server.
You are right but running Redis or any database without a firewall in 2026 is already a quite bad posture. Now with credentials as well that’s pretty much as bad as it can be.
>> Access to Redis is not access to the underlying server.
Then you should make sure you run it with the right privilege and consider it potential rogue so when it gets hacked you don’t have to worry about the blast radius because it’s contained it redis privileges.
If you are in this kind of bug hunting then try Wordpress with credentials and see how far you can go. I am sure you can find several such bugs with or without any AI.
>> Access to Redis is not access to the underlying server.
Then you should make sure you run it with the right privilege and consider it potential rogue so when it gets hacked you don’t have to worry about the blast radius because it’s contained it redis privileges.
If you are in this kind of bug hunting then try Wordpress with credentials and see how far you can go. I am sure you can find several such bugs with or without any AI.
The reason this is interesting is that it allows to go further into the target system if you already have a frontend vulnerability
It is a bug, but you've already broken into one of their data stores, which is often the goal of attackers, and you can likely bring down their service with malicious redis commands.
This is a deeply uninteresting example for anyone clueful. It’s an authenticated RCE in redis, anyone even vaguely familiar with the codebase knows to not expect there to be any real security boundary in place here.
Don’t confuse this with an unauthenticated RCE, that would actually matter. Absolutely anyone can shit out endless bugs like this with AFL, this is an extremely messy unhardened surface that expects trusted inputs.
Don’t confuse this with an unauthenticated RCE, that would actually matter. Absolutely anyone can shit out endless bugs like this with AFL, this is an extremely messy unhardened surface that expects trusted inputs.
Yup. Getting pretty tired from seeing these "0days" that are not exploitable in any real use case. Just vibes, no substance.
Tired seeing them is one thing the real problem is tired updating your software stack with a frequency the team & tooling was never designed to sustain. Worse: You are not tired enough to forget the real security issues that would need some calm period to fix.
Sad side of this is that decision makers are seeing these and think that AI can replace yet another department in their office.
I just saw a nice presentation from SecFest from a guy who ran a red team, but now they pivoted to basically anything and will be more than happy to fix your car if that would pay their bills.
I just saw a nice presentation from SecFest from a guy who ran a red team, but now they pivoted to basically anything and will be more than happy to fix your car if that would pay their bills.
Sorry but this is simply delusional. We're in an era where creds can be bought in online markets, and redis open source has no default requirement for password complexity.
If you can 'shit out endless bugs like this with AFL' then maybe redis isn't fit for purpose. But I think it more likely that it just needs hardening and an end to these laissez faire attitudes.
If you can 'shit out endless bugs like this with AFL' then maybe redis isn't fit for purpose. But I think it more likely that it just needs hardening and an end to these laissez faire attitudes.
Maybe share with us one good reason to not treat redis credentials like SSH credentials?
> and redis open source has no default requirement for password complexity.
Loudly Crying Face Emoji
> and redis open source has no default requirement for password complexity.
Loudly Crying Face Emoji
You mean, like uncontrolled keys that can be used from anywhere and have no lifecycle care? Well at least SSH keys can't be guessed.
Redis suffers from security gatekeeping, where essential features are kept in the paid Enterprise fork, and community security efforts are blocked. So you can use x.509 mutual TLS certificate auth, but the setup is entirely manual. Which means you use long lived certificates, which is worse than hard passwords.
It needs short lived certificates and pinning against endpoints, and logging of endpoints against queries, and logging auth events/origins to the SIEM. If you have that setup then post-auth compromise is more manageable.
Redis suffers from security gatekeeping, where essential features are kept in the paid Enterprise fork, and community security efforts are blocked. So you can use x.509 mutual TLS certificate auth, but the setup is entirely manual. Which means you use long lived certificates, which is worse than hard passwords.
It needs short lived certificates and pinning against endpoints, and logging of endpoints against queries, and logging auth events/origins to the SIEM. If you have that setup then post-auth compromise is more manageable.
Here's a thought: if you want something to exist, create it instead of begging a for-profit company to let you use their work for free.
So combine easy public access to uber hacker LLMs with the European CRA coming into full force next year in December. cha ching!
15,000,000 euros fines for all tech companies in Europe :-)
15,000,000 euros fines for all tech companies in Europe :-)
No, CRA doesn't cover SaaS services or open-source software, unless it is part of product (e.g. router). So it wouldn't apply here.
Dumping k3 is china's trojan horse to the Nasdaq100 and software valuations in general
I wish they wouldn't have posted this for 3 more days. Every agency would now try to suppress its weight release.
https://github.com/ianm199/valdr
I made a Rust version of Valkey/ Redis was wondering if this would become relevant.
I made a Rust version of Valkey/ Redis was wondering if this would become relevant.
Anyone know the total cost of tokens to achieve this?
Interesting finding about k3 capabilities, uninteresting 0day.
Shh, maybe wait till weights are released.. without additional "guardrails", but I'm afraid that might not actually happen
it's not actually possible to make an open model that cannot be easily jailbroken. when you have access to its entire state it's trivial to gaslight it into a non-refusal state (you can forge its responses to build up the jailbroken state).
I don't know if the sole kimi k3 provider gives that level of access atm however (where you can dictate its own responses to it).
I don't know if the sole kimi k3 provider gives that level of access atm however (where you can dictate its own responses to it).
> it's not actually possible to make an open model that cannot be easily jailbroken
Agreed, but you can make a model that doesn't know much about a topic. gpt-oss is pretty well known for not being trained on erotica stuff. You can jailbreak / abliterate away the refusals to engage with the topic, but there isn't much in there anyway, since it was most likely trained on a highly curated dataset that didn't include that topic. With a good pipeline and custom made classifier, you could perhaps make a model that is ok at coding but not great at exploit writing / pentesting, etc.
Agreed, but you can make a model that doesn't know much about a topic. gpt-oss is pretty well known for not being trained on erotica stuff. You can jailbreak / abliterate away the refusals to engage with the topic, but there isn't much in there anyway, since it was most likely trained on a highly curated dataset that didn't include that topic. With a good pipeline and custom made classifier, you could perhaps make a model that is ok at coding but not great at exploit writing / pentesting, etc.
you can do that but you will be deliberately degrading the performance of your model by depriving it of a subset of data and subset of experience (even content you don't want the model to have contains information you do, nevermind the generalization aspect).
with coding & exploit development you likely cannot even decouple the two if you wanted to, and if you can it will almost certainly cripple coding. note that reverse engineering binaries contains vast amounts of data for the model to train on to become good at bit engineering (understanding compilers, assembly, cpus etc).
with coding & exploit development you likely cannot even decouple the two if you wanted to, and if you can it will almost certainly cripple coding. note that reverse engineering binaries contains vast amounts of data for the model to train on to become good at bit engineering (understanding compilers, assembly, cpus etc).
For sure. Now that I think about it, it would be a nice research question to answer: how much can you de-couple a field that is close to another and purposefully removed from the training data. At least ICL should bring back a lot, and as you say signals about one might come from different places.
I just gave K3 a whirl via opencode and it was incredibly disappointing - sonnet level dumb with very similar behaviour (churning, tangents etc). It couldn't figure out a very simple issue between two commits.
new github workflow acquired
I would like to see the chat logs and the tooling used to run 32 agents.
This will be a busy weekend for all sysadms. This is another redis 0day, this one found by GLM 5.1:
https://xcancel.com/Lyutoon_/status/2080494539513778610#m
https://xcancel.com/Lyutoon_/status/2080494539513778610#m
What kind of utterly useless sysadmin relies on authenticated redis admin surfaces to be memory safe?
What crazy environment requires low priority nothingburger bugs like this to be fixed during the weekend?
What crazy environment requires low priority nothingburger bugs like this to be fixed during the weekend?
One where customers have their own scanners, and their unfounded panic overrides logical analysis by the engineers and admins.
We’ve had to patch plenty of stupid “security” bugs just to satisfy a paying customer.
We’ve had to patch plenty of stupid “security” bugs just to satisfy a paying customer.
I guess an llm is instructed to find an exploit, it finds the one with least resistance. And stops.
Once these are closed, they'll find more
Once these are closed, they'll find more
These big LLMs are doing a good job of finding exploits, but it does seem like it's mostly "just" a matter of loosely searching for likely places to find one, then tirelessly trying all possibilities. They are mostly finding things that a human would find it a human looked at that section of code, but humans don't have time to review the whole code 24/7. I wonder if there are more intelligent ways to scan that would find even more.