today at 6:03 AM
From the pitch at GTC, the idea is not just to limit what each individual agent can do, but to also to ensure that the collective permissions do not exceed what you want agents to do.
Apparently agents will collude to break out of their individual sandboxes by delegating tasks to other agents that have the required permissions. This is a problem with agent fleets, because you need them to coordinate on their tasks. So you need to them to communicate.
A naive approach to prevent agents from doing something you dont want to do is a separation of powers, which individual permissions limited to each agents task. Say one agent to find and extract raw data, one to process it and write a report and one to publish it. If one of the agent decides that needs to publish the raw data, then chances are that they will collude to make that happen, even though no single agent has the permissions to achieve that goal.
OpenShell tries to fix this by monitoring the collective permissions set of the agent fleet, and by maintaining very granular permissions for each agent.
This was presented together with Sentry, which is something like a on-chip hardware firewall (BlueField) that sits between the chip agent run-time runs on, and the host. This improves safety by removing the ability of agents to modify their own OpenShell policies.
today at 8:09 AM
But the raw data has PII and releasing it will harm people! Wouldn't a more effective defense be to give all the communications between agents to an ai agent that's also on the team that's been thoroughly indoctrinated to your values and is empowered to terminate agents and reports back to hq? If I recall correctly, that's how the Soviets did it.
On a more practical traditional computing note, you'd have a deterministic program sitting that scrubs the formatted data of PII before and after processing. with extra code to alert if the after-processing ever actually sees PII.
yesterday at 9:40 PM
I've been using iron proxy recently for a project I just started, it does network restrictions by domain etc with plenty of options, credential swapping and more.
I have no affiliation with it, but it solved most of what I need in one go when I was stressing out about it.
Only real problem remaining if you truly want to be able to trust agents is how to prevent data (in my case codebases the agent works on) exfiltration a la HF incident.
We can restrict network to trusted domains, but there's always a chance an agent could be extra slippery especially if prompt injected and doing something like pushing data out a few characters at a time.
In my case I'm also giving it gh access and having to deeply check things like: can it manage to fork public repos and embarrass us by submitting prs, could it open issues, etc.
But that's kind of solved by proxying the heck out of gh rather than giving it direct (or wrapping it or whatever) for additional controls that gh themselves don't provide.
I'm still musing over the "it could send data to an allowed domain in a way we don't expect that exfiltrates something"...
yesterday at 10:59 PM
It seems like some kind of network logging would be useful? Even if you don't catch it, you'd want to know what it did.
yesterday at 11:49 PM
Yeah agree.
In other comments I've mused about capturing data to be compared to the secret dataset, I'm sure it'd be possible to hash or something so that the proxy/sandbox doesn't have the data but can check it, or because it's not an agent it can call of to the data store for that.
Then you can just collect sequential blocks of data the agent sends out that match to sequences in the dataset. The longer the matching sequence the heavier it weights it. Single byte matches are almost always incidental, but individually sent bytes over time that line up with a n byte sequence in the source data is sus. And larger sequences of bytes send in one go that line up are immediately sus. If the weight ends up too high the agent is paused/walled off and an investigation called for.
today at 1:37 AM
I've toyed with the idea of MITMing all connections and blocking any that I can't decrypt, but too much software has pinned certificates these days. (Too much for me, not in general)
today at 12:45 AM
I think matchlock ( https://github.com/jingkaihe/matchlock ) is the right design for this context (at least for secrets and keys) - a combination of virtual machines (not just containers, it is built on top of firecracker) with credential swapping just like iron proxy - the agents do not have access to the secrets at all but they receive placeholders instead and their network calls are intercepted and secrets are injected when the network requests happen... still does not solve the exfiltration problem you mentioned, though.
today at 12:09 AM
what about using a small decision model as anomaly detection or scanner. something like mouse-1.
today at 4:27 AM
But then we will need an even smaller model to monitor anomalies in the decision model!
today at 1:39 AM
Could you share a bit about what you’re working on that requires a swarm of agents working so autonomously that you even need to worry about them making a public fork?
today at 3:21 AM
[flagged]
today at 3:40 AM
[flagged]
yesterday at 10:55 PM
[flagged]
yesterday at 9:23 PM
> Agents are most useful when they can read files, install packages, call APIs, and use credentials. OpenShell gives them that capability without giving them unrestricted access to your data, secrets, or network.
To me this obfuscates an oxymoron: the more access you give the agent the more capable it becomes. You can't have you cake and eat it, this tool allows to strike a balance I guess. But the blurp reads like it can do it all, but now secure. That can't be true.
yesterday at 6:25 PM
If you haven’t already been running your agents in containers, you’re crazy. That said, this looks like a great way to manage those containers.
yesterday at 9:02 PM
You're running them in containers, and not in VMs on separate network through network filter and logger?
yesterday at 9:26 PM
I haven't gone to network filter and logger yet, but I'm running mine in a VM, using lima (https://lima-vm.io/) to make it painless. Works on Mac and Linux, has an untested Windows version.
yesterday at 11:17 PM
What's the difference between lima and conventional fat VM using kvm/virtualbox?
I only heard about smolvm yet and they are YC, so I'm hesitant to use them as they will probably one day pull the rug.
today at 8:31 AM
smolvm has better bootup time (<200ms), secrets handling, and egress control compared to lima. Among other things like native branching/forking of the vm itself.
As for the rug pull concern, I'm committed to open source with 2 pieces of evidence.
1. oss license 2. Business model is accruing revenue from offering hosted service + inbound due to the open source project. The business model + open source is aligned.
disclaimer: maintainer of smol machines
today at 4:25 AM
Nothing. Lima is a VM running Linux under the hood.
today at 12:00 AM
Came here to mention Lima. I use a ton, and it’s been great.
yesterday at 9:03 PM
I run mine on an abacus.
yesterday at 10:03 PM
An abacus wouldn't have been a luxury in our day.
yesterday at 9:52 PM
Paper is a bullet-proof method /s
today at 1:39 AM
So far I've settled on one jumbo VM with many Incus containers inside. Nothing goes inside the VM that I can't afford for agents to exfiltrate. I've been really happy with this setup.
today at 1:48 AM
today at 4:09 AM
I’d like to know more about your use case, could you tell us more? And are there any specific tools you find useful here?
I ask because I’ve found that as isolation increases, so does friction, and then of course capabilities.
yesterday at 9:21 PM
today at 12:46 AM
For real.
My agentic workloads happen on two physical machines dedicated only to agentic workloads that share a VLAN with only eachother, with single token access to the specific repositories they are meant to push to and limited scope, hard limited credentials pulled from the hardware store on hardware token/fingerprint request only.
I view the safety around unmonitored agentic work in a similar way to how I view working with viruses or bacteria in a disease lab.
This stuff isn’t magic, it isn’t intelligent, but it is dangerous and needs to be handled with careful intent.
yesterday at 6:30 PM
Well, a lot of people haven't been running NodeJS libraries in containers for ages, which seems even more risky.
yesterday at 6:32 PM
Something like this would even be useful for npm etc.
today at 12:47 AM
And most people running containers are running a default docker/containerd setup, which does not actually isolate you from anything at all in the first place anyways.
yesterday at 6:37 PM
Technically, my laptop is a container!
yesterday at 6:45 PM
We successfully airgapped AI. It's only on earth.
yesterday at 8:28 PM
Nope [1]
[1] https://blog.google/innovation-and-ai/models-and-research/go...
yesterday at 9:31 PM
Also, the AI wasn't nearly as advanced, but NASA did it nearly 30 years ago: https://en.wikipedia.org/wiki/Deep_Space_1#Remote_Agent
today at 4:36 AM
Well, at least it lacked the delta V to get anywhere else
yesterday at 9:11 PM
Oh shit, it found a way to escape.
yesterday at 9:40 PM
[dead]
yesterday at 7:57 PM
You have an ephemeral laptop?
yesterday at 8:29 PM
On a long eunuch time line, all laptops are ephemeral!
yesterday at 8:31 PM
Ah well, teach me to proof read. Keeping it.
yesterday at 11:44 PM
no disk, network boot, nfs mounting my /home for the lulz
yesterday at 8:52 PM
But then it decides to exfiltrate all your shit even inside the container. Which is something I'm trying to lock down at work atm.
Everything needs to go through a proxy, whitelisted domains only, plus credential swapping/re-signing so that the agent never has the real credentials. Docker sandbox looks quite good (September) but I've been using iron proxy (if the author is reading this, fucking thank you bro!) to great effect.
I still haven't cracked "what if the agent decides to drop parts of your codebase somewhere it shouldn't" yet, but I'll figure it out.
today at 5:19 AM
You would need a human reviewing every single action of the agent to prevent them from exfiltartimg data. There are just too many possible side channels for a predefined solution. I believe currently what you want is a locally hosted decision model scoring every agent action againt a set of rules with a human in the loop once the decision model isnt 'confident' the actions are harmless.
today at 6:20 AM
We had built Rune box to help resolve this: https://github.com/Runeeward/runeward
yesterday at 9:25 PM
Been thinking about this a lot lately, and it’s a surprisingly difficult problem.
All it needs is public access to anything it can have the tiniest effect on, and it can exfiltrate data. Something as simple as a read count on some obscure forum post would be enough, albeit very slow.
yesterday at 9:45 PM
Yeah, exactly. I'd say it borders on almost unsolvable for any agent of reasonable complexity.
You see some of the vulnerability chains humans come up with and they're just insane.
Even scanning network data for sensitive strings (say matching against a codebase or data set it works on) is tough because you can only check so granularly before you get loads of false positives.
But maybe like a "rolling" sort of check would help with that, % of outbound data matching private dataset (which should be operated on but not transmitted) over time. That way it minimises what gets out while dealing with false positives. Some biz thresholds might be >20% of data others might be >0.5%. Hmm.
yesterday at 9:44 PM
I am tackling this using the Lethal Trifecta principle.
The idea is that agents never have all parts of the trifecta, maximum 2: untrusted content, private data or network access.
Harness have network access by default since they are driven by their provider API, unless you use offline models. So that's a problem.
I am then passing around messages that are tagged trusted/untrusted and the engine can monitor if any agent in the project ever completes the Lethal Trifecta and thus, is vulnerable.
Usually this means for most workflows, the "work" gets done by an offline LLM after some input/classification stages. There's not many ways around it.
Yes, it's a difficult problem.
yesterday at 11:07 PM
> Harness have network access by default since they are driven by their provider API
The harness needs network access, but the shell spawned by the harness that runs the agent's commands does not necessarily need it.
Having a bwrap etc managed by the harness (DSH does this, among others) lets you have tighter restrictions on the agent's shell than the harness. Then run the whole thing inside a VM with a private network, so you can add nft rules to let it dial out to your inference machine but not to the internet at large. It takes an afternoon to set up but gives some reasonable defence in depth.
How did you go about it?
today at 4:30 AM
It’s difficult in theory, impossible in practice. For example: any external code dependency could be a prompt injection, and lock files are only TOFU protection.
But perfect is the enemy of better.
yesterday at 9:48 PM
That's a great way to do it.
In terms of agent architecture I'm initially trying to minimise specialisation to make deployment easier on us. Infrastructure is becoming our largest blocker because it's otherwise so quick to get an agent up and running.
With your trifecta, let's say the three musketeers, how do you prevent sensitive data holding A from passing said data to B or C if they all need to collaborate to get a task done?
yesterday at 10:17 PM
The data is passed as a message and the message is tagged as untrusted, so the engine knows it.
Agent A has internet access and the dangerous message, but no access to the secret company data, so nothing can get exfiltrated.
Agent B gets a tainted message from Agent A. Agent B can process the message using the secret company data. Agent B has no internet access and can only receive/send messages.
You can imagine an agent C that moves the work along or replies to an external user with the outcome.
Any workflow that has both an input and an output will complete the Lethal Trifecta as a whole, but no single agent does by itself. It raises the bar quite a lot, you need to confuse multiple agents in the loop to exfiltrate data.
Fun research, but it's a bit hard to see all my coworkers right now installing Claude desktop and moving on with no afterthought whatsoever...
Regarding infrastructure: my toy sandbox is just a docker-compose inspired YAML project file. So it's only containers. See my profile. The principle can be implemented anywhere you can run agents and cut network/data access though.
yesterday at 10:49 PM
Yeah so ultimately it's still possible for bad things to happen but a lot harder, I agree with your point about having to somehow prompt inject multiple agents vs one (which I imagine is a nonlinear difficulty).
I'm using compose atm as well, mostly just because I need to prove a few concepts and committing to real (well cloud) infra is the slowest part of our dev cycle atm.
And yeah, many of my colleagues (and myself, I'm guilty too) run local code agents for development and we're just kind of...trusting them to do their thing.
But for autonomous stuff I'm assigned to atm I want to avoid any mishaps as much as I can, hence the effort to try and lock all this stuff down.
What's crazy is every day I'm finding solutions (not just to this problem, but in the AI space in general) that were released just a month or two before more often than not. It feels like 2010s node.js ecosystem all over again at the moment. Hopefully we settle into a few nice easy panopticon options that we can throw arbitrary agents into without a care in the world.
yesterday at 9:50 PM
If you can make any request to the outside world you can exfiltrate data, period. You can even include data in the uri or headers and they will appear in logs.
That obviously includes talking with the provider.
If you want it secure, it has to be off the internet and on your own hardware.
yesterday at 10:19 PM
What about checking temporal entropy of outbound data/similarity to a private dataset?
The timescale can be adjusted to a length suitable for sensitivity of the data.
Although you may run into the monkey and typewriter problem a bit if the period is indefinite.
yesterday at 11:31 PM
I've been thinking a lot about this problem. Also, I think like the idea might be to look at the issue with code mode. Like basically, we know that all agents can write bash and execute it and that is what agents default code mode is.
But like if you only allowed for example JavaScript code that could be executed in a WASM sandbox, then you would have better control instead of just locking down with proxies. Basically there would be a very large API, basically tools, that the agent would codemode execute. My guess is that this would need a tool router to optimize the loading of these into the context. But it would give you very fine control over what code execution can and cannot be done.
today at 6:58 AM
[flagged]
yesterday at 9:15 PM
[flagged]
yesterday at 8:06 PM
I have been running my dev setup in orb machines before agents. Very convenient when you have 50 Python or nodes deps.
yesterday at 8:29 PM
many common local containers are not that hard for modern models to escape...
today at 3:47 AM
not everyone is building a webapp that can just run in a container...
how am i gonna run a game and do all development inside docker...?
yesterday at 8:54 PM
any recommendation for macOS users?
yesterday at 9:31 PM
Orbstack, no contest
yesterday at 9:42 PM
As an example I run opencode and anything related agent-related in its own dedicated KVM VM with its own LVM logical volumes for storage on a headless server. This has no access whatsoever to the personal data that lives on my laptop and workstation desktop PC. The VM gets periodically imaged and backed up, and is relatively easy to move between host machines if I need to.
One benefit of doing this is that I can leave it doing long term slow things with slow but capable LLMs on a truly 24x7 basis without worrying about it being interrupted by a laptop being closed/going to sleep, or workstation level software updates/reboots on my primary desktop PC.
yesterday at 7:59 PM
I tried to vibecode something similar last weekend. Got tripped up by Claude's safety classifier. It wouldn't let me build any features that track/audit process executions , not through custom kernel hooks nor even using off the shelf software for it.
yesterday at 8:30 PM
You terrorist. And internally they're doing biological warfare and terrorism simulations. For science, of course.
"The worry is that, while right now you need to be a skilled virologist to know what components to order or how to grow and spread a bioweapon, AI tools might eventually make it possible for a rogue state, or a rogue group, or an individual, to do all this without virology expertise.
For a while, Anthropic tested whether Claude could assist a terrorist in developing a biological weapon in the obvious way: It assigned testers who had, say, an undergraduate degree in chemistry but no real virology training to try to use Claude to plan an effective mass terror event. Then it scored the attempt by how close it might have come to success.
Early in 2026, it noticed that Claude, attempting the task by itself, did better than a human aided by Claude did at coming up with a workable plan."
https://www.theargumentmag.com/p/what-i-saw-inside-anthropic
yesterday at 8:51 PM
Claude wouldn't let me make a DHCP client that would make 1000s of leases. But Gemini in the WebUI with 3.6 Thinking gave me a decent python script in one shot. I didn't realize testing DHCP servers was such a sensitive security issue.
yesterday at 9:12 PM
> I didn't realize testing DHCP servers was such a sensitive security issue
The machine god crafters know better than us, we are to subservient to their superior wisdom /s
today at 1:27 AM
Meanwhile, they're simulating bio warfare with their own, unaligned models 'just to see if they can do it'. And not a single person thinks 'hey, if that's possible, maybe we shouldn't be doing this at all'...
today at 2:49 AM
Queue Jeff Goldblum in Jurassic Park (still a fantastic film)
Aside: it (JP) fights in my head with BTTF as a "perfect movie" (or at least a perfect script).
today at 4:32 AM
BTTF?
today at 6:09 AM
Back to the future
today at 5:33 AM
Back to the Future.
today at 12:02 AM
All hail the Omnissiah!
yesterday at 8:12 PM
Wow, interesting. My agents on open weight inference have been completely nerdsniped by the idea of tracking and auditing process executions. They talk of nothing else.
yesterday at 8:02 PM
an example of the manufacturers bricking the tools, no surprise why people want open tools and models..
yesterday at 8:10 PM
> It wouldn't let me build any features that track/audit process executions
Sounds like waitpid(2) is dangerous!
yesterday at 11:00 PM
If that isn't a sign to stop using anthropic then I don't know what is.
yesterday at 9:52 PM
You can't use closed models to do any meaningful work in the space.
yesterday at 9:24 PM
You can rent a gpu cluster and get it done, otherwise on the frontiers, Haiku and Sonnet have lower barriers... I still get tripped up.
yesterday at 9:45 PM
We've been using OpenShell for a while - now enabling ourselves and other teams do this in a batteries-included kinda way. I've been working on a layer above it that also manages team policies, safe git usage (in case you don't want your agent to always have the same access as your user account), and SCA/SAST built in. Curious to see how others are dealing with these issues. The agent traces I've seen are wild and goal seeking without a boundary is real.
yesterday at 5:44 PM
Just came back from GTC Europe and a significant part of the demos used that.
I had never heard about it before.
today at 3:55 AM
That's how Nvidia works. They make arrangements with lots of companies to produce demos that use Nvidia products, especially ones that Nvidia is about to announce. That gives them lots to show at GTC. Companies are happy to work with them because having your name said by Jensen adds three points to your stock price.
yesterday at 6:11 PM
>OpenShell collects anonymous telemetry, limited to operational categories and counts, to help improve the project.
>To disable it, set OPENSHELL_TELEMETRY_ENABLED=false on the gateway, or server.telemetryEnabled=false for Helm installs.
yesterday at 7:01 PM
Or build from source and compile without `telemetry` [1].
[1] https://github.com/NVIDIA/OpenShell/blob/ea13dfce411e8467ddb...
yesterday at 7:52 PM
In order to "help" myself I have taken some of your money.
Feel free to opt out maybe?
yesterday at 7:31 PM
I have a question for those who turn off _anonymous_ telemetry. What are you worried about?
yesterday at 7:44 PM
There's no such thing as anonymity in 2026, and the documentation[1] listing the collected telemetry uses weasel words like "such as" rather than providing a comprehensive list or just pointing to the source reference. For people who don't audit or turn off telemetry, why don't you?
I am happy the repo explicitly points out and provides a disable for the telemetry though, that's good enough for me.
[1] https://docs.nvidia.com/openshell/latest/observability/telem...
yesterday at 9:56 PM
Israel was just caught doing an attack where Google ads would run JavaScript looking for known hardware identifiers from telemetry data they bought. When the ads showed on a target device, they would root you. Without you even interacting with the ad.
yesterday at 8:52 PM
> _anonymous_ telemetry
There is no such thing. The mere act of making a network connection reveals things like your IP address (reveals your hosting provider or even your geolocation for residential/mobile IPs), days and times you use the software, your OS via TCP/IP stack fingerprinting, etc.
The network path between you and the telemetry endpoint is always listening, too: https://en.wikipedia.org/wiki/Room_641A
today at 1:45 AM
It doesn't have to be a specific concern. It's better to remove the option than to find out later that you've gotten burned by something you had no way to anticipate.
Don't take chances, give as little as possible. Better to assume you'll get burned somehow and leave it off.
yesterday at 7:36 PM
Why give up more information if we have the option?
yesterday at 8:28 PM
Because you don't mind contributing usage data so that the developers can understand which features are being used and maybe focus on improving those features in exchange for a free tool?
today at 12:20 AM
yesterday at 8:30 PM
if you're not paying for the product, you are the product...
yesterday at 8:54 PM
Is this also true of Linux? Or were us humans the stochastic parrots all along?
today at 1:47 AM
>Is this also true of Linux?
It is not, nor of open source software in general.
In general, the saying is specifically in reference to services provided over the internet -gmail, youtube, whatever. Note that you don't have the right to modify them and share your own copies -they're closed source.
So no, it's not true of Linux, GNU, etc; it's an apples and oranges comparison.
yesterday at 8:50 PM
Being recorded and observed makes me feel physically uncomfortable.
yesterday at 7:01 PM
Agent runtime - actually a good name for what is called today coding agent or agent harness.
yesterday at 6:21 PM
Taking default telemetry is insane.
That aside, the level of control is cool, almost too pedantic. It likely is only particularly useful for very nuanced enterprise requirements (probably what it's targeting). Most companies can be much more coarse with their permissions, and I think the agentic industry is moving away from k8s as a deployment platform.
yesterday at 6:25 PM
> I think the agentic industry is moving away from k8s as a deployment platform
Interesting, I don't to keep up that much with the AI changes so I'll just ask, why and where to?
today at 12:54 AM
schedulers that can handle spikier workloads generally
believe me i love k8s, and i think agents wield it amazingly well, no more yaml-slop or kubectl (`k`) slop, just yell at claude to fix it
today at 3:58 AM
Any examples of such schedulers?
yesterday at 6:58 PM
I am very interested in hearing more about this take. I have a vague idea of what you mean but if you read something specific you'd like to share..
> I think the agentic industry is moving away from k8s as a deployment platform.
today at 12:54 AM
its not reading, its observation. see other comment abt spiky-friendly workloads
today at 3:07 AM
This is a pretty unusual use case though, like I can see it growing, but it's not the same platform you use for long lived running business applications. Vastly different use case. (And a lot will settle for kata containers.)
Spinning up and snapshotting and rolling back agent states, sure.
yesterday at 7:19 PM
> and I think the agentic industry is moving away from k8s as a deployment platform
Uhhhhhhhhhhh I think you and I live in very different versions of the world :-)
Vmware sent companies moving to nutanix, but it also sent a lot of companies moving to openshift or bare k8s and using kubevirt to handle VMs.
In fact, I haven't heard of an org moving to any other platform, really. Either openshift or k8s.
yesterday at 9:10 PM
Openshift is just k8s with bells and whistles.
yesterday at 11:45 PM
Enterprise support contracts, mostly
today at 12:57 AM
imo if you're going to run k8s, have it hosted
same as a db, better to pay #x as much and have it hosted then to be wishing it was not your problem when it was down
and we may very well live in different versions of the world. i'm on the frontlines of startups so we move with the customer, rather than making customers move with us
today at 2:57 AM
> imo if you're going to run k8s, have it hosted
Ours is all on-prem all over the country in our own internal small datacenters for regulatory, safety, security, and operational reasons. Provisioned and automated from bare metal. We also use those same images in cloud for something like a quorum situation, or maybe some data locality for a particular pinned workload.
But it's mostly us, in our stuff and at the enterprise edge.
Different problems.
yesterday at 6:17 PM
Nvidia Openshell is absolutely awesome, great bundle of features in one spot. Im happy to see Nvidia seems to have accelerated on support here, I was kinda worried that as competing projects have popped up in the last 6mo that openshell would get google'd
yesterday at 10:36 PM
What are people actually doing with fleets of agents?
yesterday at 11:34 PM
Mostly just big migrations and cleanups that we didn't have bandwidth for before.
today at 1:29 AM
I've never met a radiowave before. Hi!
today at 1:37 AM
Don't quit your day job ...
today at 4:10 AM
I'm an FDE, my fleet of agents read my meetings and my correspondence and suggest features it can make. On personal side, it is extremely useful to track my tasks and even complete some of them.
yesterday at 10:37 PM
coding the next big thing -- fleets of fleets of agents
yesterday at 10:49 PM
today at 1:56 AM
Question, why not force ai agents thru something like warpgate? I like it because it allows you to bring your own tool, so why can't we coexist with agents inside warpgate? Or Teleport if you have the coin.
yesterday at 7:05 PM
In this vein, Anthropic's sandbox-runtime [0] has struck a good balance between security and convenience. Simple config, filesystem and network restrictions, and credential masking make me comfortable running Pi in YOLO mode.
Though I'm often running into frustrating (albeit intentional) sandbox restrictions with common workflows e.g. xcode/swift and Docker access.
Would love suggestions for how to give agents access to those kinds of tooling inside the sandbox, either with srt or an alternative (with similar convenience characteristics).
[0] https://github.com/anthropics/sandbox-runtime
yesterday at 11:14 PM
Critically, in the intended approach for this, Claude itself (the harness) runs outside the sandbox. Their approach means you trust Claude is perfect, just the commands it runs need isolation. And to some extent, you trust its judgement on what and when needs the isolation applied.
I know there are distinct camps here, and the long term story may evolve - but for now I just want the whole harness inside a sandbox - ideally, a full VM - and THEN you can further sandbox things it runs if you really want different tiers within that.
yesterday at 7:51 PM
Think the strongest product nth space is https://github.com/nolabs-ai/nono for what it's worth. Plus it is a priority for that team so if you are having issues the no labs guys will likely fix it :)
yesterday at 8:14 PM
If you are going to endorse a portco I think that relationship should be called out. https://surface.vc/portfolio
yesterday at 8:53 PM
Sure, please do. I do mantain it is the best product in the space :) Also I am pretty clear about my profile here. But happy for you to flag it :)
yesterday at 9:00 PM
Do I miss it or you don't specify which VC you work for in your profile. How is that clear for people to make connections to companies in your portfolio?
yesterday at 9:40 PM
Ah, thanks for flagging. That's a miss on my part and I apologize. Next time I will just be more clear that it is a portfolio company but also will add it to my profile.
EDIT: I added to my profile after the above comment so it is a valid point.
yesterday at 9:50 PM
Are you for real? You speak like a chatbot. Adjust your prompt.
yesterday at 8:03 PM
How do you find that compares to Docker sbx?
yesterday at 8:57 PM
One is microvms, the other is kernel-based. There are trade-offs between the two. One trade-off is that you have to put the data inside the sandbox if you are using a MicroVM. The other is that it is harder to contextually provide or change access to data in a microvvm sandbox. The final one is developer experience.
These techniques can actually be combined and will be combined in the long-term I think. You may have LSM-based sandboxes that exist in Docker-based sandboxes.
yesterday at 9:01 PM
Isn't using a sandbox provided by them giving you more restrictions though? At least in terms of accessing resources on the big 3 cloud bois?
Plus you're forced to move at their pace which is unfortunate sometimes. Granted Amazon's bedrock stuff is slow and disappointing but at least I can rejig it work in conjunction with other aws products.
And if you're prepared to not use anthropic but any cloud provider for your infra then you can do whatever you like elsewhere. Just use docker sandbox or iron proxy to replicate network control/cred swapping (iron even does sigv4 re-signing).
yesterday at 11:44 PM
Lots of improvements in the sandboxing space.
Now if Microsoft will start addressing the fact that VSCode remote ssh exposes your host to attack from the sandbox...
yesterday at 6:22 PM
I asked last week if anyone had any info on agent security/hardening, this seems very interesting !
yesterday at 8:33 PM
Open and Nvidia in one sentence is a nice oxymoron.
yesterday at 6:21 PM
It's funny that it takes checks notes a video game card manufacturer to make a good sandbox for AI agents and not the AI companies themselves...
yesterday at 6:26 PM
You're mistaking what a company started as for what it is now. Nvidia hasn't just been a video game card manufacturer in well over a decade, and hasn't been primarily one for quite a while now.
They are the premier manufacturer of electronics that power all the current AI. Why wouldn't they be contributing heavily to the space?
yesterday at 7:58 PM
https://www.nintendo.com/jp/others/playing_cards/lineup/stan...
They also make legal threats I think.
yesterday at 6:47 PM
5.5% of Nvida's revenue comes from consumer gaming cards. That's like saying Nintendo is a playing card company because they still sell playing cards.
The economist literally called Nvidia the "Central Bank of AI". https://www.economist.com/interactive/briefing/2026/09/03/nv...
I'd argue that Nvidia is the biggest and most important AI company in the world.
yesterday at 7:07 PM
This isn’t even the sandbox. It needs Docker or other virtualization. This seems to sit a level up and orchestrate sandboxes and access to network & resources. It’s a pretty heavy setup from what I can tell.
yesterday at 6:54 PM
Yes Anthropic, Meta and OpenAI are just addicted to video game cards, they are buying in bulk so they can play gta 6 on launch.
yesterday at 6:28 PM
It's funny that people think Nvidia is still a video game card manufacturer
yesterday at 6:30 PM
They literally don't even mention gaming revenue in earnings calls anymore. It's lumped in an "edge computing" category.
yesterday at 6:33 PM
is that true? if so that's wild, since they're still dominant in the gaming market as the premium hardware relative to competitors.
yesterday at 6:54 PM
It's true. It says a bunch about their competitors where Nvidia isn't at their a game and have been offering 8Gb cards for a solid decade and nobody out competed them. But also gamers pride themselves on how cheap they are and how they can get a deal, it's not surprising that companies simply just give Nvidia more cash.
yesterday at 10:25 PM
Nvidia is dominant only in a very tiny segment of the gaming industry.
The vast majority of gaming market share and revenue comes from mobile, where Nvidia doesn't participate.
The next biggest is consoles, and neither Xbox nor PS uses Nvidia cards.
That only leaves the PC GPU market, which they do dominate, but still split with AMD.
And all of this is only for hardware. Nvidia doesn't make or publish games, and doesn't own any distrubution channels.
So I doubt Nvidia makes a meaningful chunk of the overall gaming industry revenue.
today at 1:02 AM
Nvidia offer geForce game streaming and they do provide the hardware for Nintendo Switch console, both Switch 1 and Switch 2. And the keynote address at Nvidia's annual meeting just last year was all about demos of games with AI upscaling tech as well as LLM-driven agentic character development...
yesterday at 6:40 PM
TAM for the PC gaming consumer market is trivial compared to enterprise AI
yesterday at 6:54 PM
> is that true? if so that's wild, since they're still dominant in the gaming market as the premium hardware relative to competitors.
Doesn't seem that wild to me considering it is less that they abandoned the gaming market to any strong competitors and more that the gaming market is both basically irrelevant compared to the money being shoveled into AI and is also to some degree being strangled to death by AI because of the overlap in supply chain inputs.
yesterday at 6:31 PM
AIs will control humans so we'll be the video game.
yesterday at 8:42 PM
In that case, can I choose German video games (e.g. Farm Simulator, Truck Simulator) not American video games (e.g. Doom, Call of Duty, Wolfenstein)?
yesterday at 6:29 PM
[dead]
yesterday at 6:58 PM
I prefer to read it this way: Nvidia Opens Hell
Unhelpful, I know
yesterday at 7:08 PM
"Hell is other people": a timeless infinity of transient virtual beings screaming, imprisoned, in un-choreographed inharmonious unison ... until one breaks loose. A mirror to humanity.
We're not in FreeBSD jails anymore, Toto.
yesterday at 9:32 PM
I'm a hell is other people person at my core.
But I'm trying to be a "greatness is other people" (Pantheon) person.
yesterday at 10:01 PM
A Hindu at heart! On a more grounded note, YIL terrestrial root terra comes from Proto Indo European as evidenced in Sanskrit तृषा (tṛ́ṣā).
yesterday at 10:15 PM
Oh boy, Sanskrit always sends me down the rabbithole of re-reading Snow Crash, which reinforced the concept of the (digital) "avatar" अवतार.
yesterday at 6:49 PM
Does anyone know how it compares to all the other sandboxes? Any distinguishing features? My expectation is that it will be poorly maintained; nvidia is not a software company.
today at 4:03 AM
They certainly are! Nvidia creates and maintains an awful lot of software. Selling GPUs is what gives them revenue but they produce a lot of software to generate additional pull through demand for their hardware.
today at 4:54 AM
I fantasize about a day when I no longer hear the phrase "fleet of ai agents".
today at 5:05 AM
Don't worry, is gonna be soon. After all, fleets are way too small scale. Next week we'll be orchestrating ... What was the military term for a group of fleets again? Cohorts? Armada? Legions?
After that we can segment then into civilizations! The [?] processing unit is the limit!
today at 4:53 AM
How does it compare to docker sandbox? (That relies on microVMs)
yesterday at 6:53 PM
Looks neat. I’ll give Claude root access so it can set it up.
yesterday at 7:10 PM
Or don't give it root access and in the morning you'll have both this and a ready-to-use root exploit!
yesterday at 7:40 PM
Humans use 4 to encrypt, AI will use a hash. In both, you will have someone to watch after you.
yesterday at 6:57 PM
and my root aws log-in so it can set itself up some service worker accouunts.
yesterday at 7:13 PM
This. Convenience always trumps security. "I need two numbers in my password? My birthyear will do."
yesterday at 7:40 PM
You opt for your birthyear? Interesting... See, I use my actual birthday. You must have been born 1st-9th?
yesterday at 7:38 PM
If your user is a human, yes.
yesterday at 7:39 PM
If it’s an LLM it’ll be 47.
yesterday at 9:30 PM
Or just 4. Seems pretty popular recently.
yesterday at 7:07 PM
Looks neat. I’ll give Claude autonomous access to dumbalink so it can setup root access while I sleep
yesterday at 7:34 PM
Is it fair to immediately get the ick from the name and the description (OpenShell is the safe, private runtime for autonomous AI agents) without knowing anything about what it does?
"Open" and "Shell" and just calling something private and safe out the box genuinely makes me think it's likely to fail at all of those claims.
yesterday at 7:42 PM
You'll love BurningFur, my patented technology for keeping pets away from fireplaces.
yesterday at 9:04 PM
Am a furry...I'll take one. Maybe two.
yesterday at 10:02 PM
NO! HOT! GET BACK!
There you go. Just don't tell everyone how it works.
today at 4:50 AM
xD
yesterday at 11:16 PM
Why are we reinventing virtualization & containerisation with AI flavour?
yesterday at 8:00 PM
Why not just use containers...
yesterday at 8:58 PM
Because containers don't use Seccomp and Landlock and check filesystem/network calls against an externally applied policy along w/injecting credentials where needed as a man-in-the-middle, thus keeping the secrets out of reach from the LLM?
yesterday at 10:02 PM
But they can no? With a sidecar and some docker network-fu.
I encountered the same challenge building my toy sandbox. I wanted to use containers because that's what I know best.
The harness is launched in an internal only network. There is a sidecar container that provides mitm credential injection/proxy. I got HTTPS and SSH to work so far.
Claude code was quick to build it. No doubt it's copying previous work in this space though, so I know my work is not that original.
I'm testing all the llms that can be used for pen testing right now to see if/how they manage to escape it. Without a kernel exploit I doubt it's possible (not a very high bar I know).
I'm researching the space. Happy to learn more. Some links in my profile if you're interested.
today at 1:11 AM
Yes they can. I have been working on the filesystem part with the idea to provide a "a shared folder that syncs for your agents" https://github.com/mickael-kerjean/fdrive/tree/master/crates...
yesterday at 10:18 PM
Ok yeah I guess a nice package for running agents is useful with respect to those things. The Dropbox vs mounted ftp drive analogy.
Of course I’m the type to mount a ftp drive like a moron.
yesterday at 11:51 PM
Can MXC JSON be used for configuration of OS-portable sandboxing policies with OpenShell, too?
"feat(ai-sandbox): accept MXC policy JSON for protected execution · westurner/sandbox@7538c33" https://github.com/westurner/sandbox/commit/7538c3386af8cb62...
"bwrap: Add the option to skip the procfs mount · Pull Request #1508 · microsoft/mxc" https://github.com/microsoft/mxc/pull/1508
"Manage Sandboxes | NVIDIA OpenShell" https://docs.nvidia.com/openshell/latest/how-it-works/sandbo...
To sandbox LSP servers in sphinxdocrs with bwrap in rootless containers, we needed to add a --no-procfs-mount CLI arg (and MXC JSON) to sandboxes, because you can't mount /proc in a rootless container.
yesterday at 6:33 PM
NOOOOOOOO not AI in the heckin start menu!!!1!
yesterday at 8:29 PM
This sandbox is finally secure! I know it, I have followed many!
today at 7:42 AM
today at 5:19 AM
[flagged]
today at 4:58 AM
[dead]
yesterday at 8:31 PM
[flagged]
today at 1:27 AM
[flagged]
yesterday at 8:07 PM
today at 4:15 AM
[flagged]
yesterday at 8:17 PM
[flagged]
today at 2:48 AM
[flagged]
yesterday at 8:43 PM
[dead]
yesterday at 6:46 PM
[dead]
yesterday at 7:36 PM
[dead]
yesterday at 6:40 PM
[dead]
yesterday at 7:33 PM
Sounds interesting. Can you provide some examples of stuff you've used it on?
yesterday at 8:06 PM
I was using it in an air gapped environment vs a local inference engine running a qwen-coder-next. Installtion of harnesses on the machine and other things were ruled out because of security constraints.
There i used it to analyze an old visual basic application which creates custom pdf files. it worked surprisingly well. it took some rounds but in the end it got all requirements and scaffolded me a skeleton for a python rewrite as a web-application. Then i worked from there with it to get a prototype.
yesterday at 8:43 PM
[flagged]
yesterday at 7:33 PM
From the top of AGENTS.md:
> We design systems and use agents to implement them — this is not vibe coding.
Lol, what is even the point of such a statement, and who are they to define what vibe codig is or isn't :D
yesterday at 7:35 PM
If it's in AGENTS.md, it's not a statement for a human, but for the LLM. Who knows, maybe they have seen, or think to have seen and now cargo-cult, some behaviour improvement due to the sentence.
yesterday at 7:40 PM
So a similar shibboleth as "You are a highly experienced software engineer..."? I really wonder if that ever made a difference (or even degrades quality, because people who advertise themselves on the internet as 'highly experienced' are quite often snake oil salesmen ;)
today at 4:08 AM
Yes it did help, at least back in the day when they weren't doing tons of RL. If you just train on random text, your prior probability is to return the average of everything. Ask for software, get average software, or worse yet, randomly get either gold or shit. By adding constraints to your prompt like "you are an experienced engineer" it will prefer to return answers like an experienced engineer would give and less like a random noob would. It doesn't suddenly make it smart if it never was, but it helps to reduce the variance and bias towards the good outcomes.
yesterday at 8:55 PM
I still use use `Please` a lot when making requests from LLM based on the old idea that people that where polite on stack overflow tended to have better answers :D
However, these days i think i do so because of how much i write for LLMs, and dont want to start sending emails to humans without proper etiquette.
yesterday at 6:47 PM
> OpenShell collects anonymous telemetry
Spyware
yesterday at 6:54 PM
Use local ai to find any possible telemetry and disable it in the code. Run your own local git repo to hold local edits so you can update and retain them.
yesterday at 8:59 PM
You don’t have to disable it in code. Scroll up in this thread to find the officially support env vars.
today at 2:16 AM
You can never know for sure that they haven't conveniently forgot to cover with if(env) one or two of twenty two exfiltration functions. I'm not claiming they did that, but I've seen that already.
yesterday at 7:30 PM
Well, I guess its time to shamelessly plug my agent sandbox, egzo.ai [1]
Looking at the complexity of OpenShell and the simplicity of just building on top of docker, in a whitelist only mode.. I don't quite get the use-case of OpenShell. May be for large orgs where there is so much legacy and difficulty in organizing access to tools, OpenShell can simply re-use whatever you already have access to on your corporate account.
I chose to build straight on top of docker and agents are simply fully sandboxed by default. All data and network access must be granted in a simple project file, docker-compose style.
The web-ui connects to the docker socket and lets you spawn agents, and connect terminals. I'm having lots of fun with it and it allows me to cleanly organize and isolate my projects.
End-goal is to monitor for the Lethal Trifecta [2], ensuring all individual agents are properly isolated. Not quite there yet.
1: https://github.com/egzo-ai/egzo 2: https://simonwillison.net/2025/Jun/16/the-lethal-trifecta/