Rendered at 06:53:37 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
docheinestages 15 hours ago [-]
I think the most secure setup, though not so convenient for the average user, is a separate machine with QEMU/KVM. The machine should be isolated adequately, such that even if compromised, it shouldn't be able to cause damage or gain access to other machines. Additionally, a proxy server on your machine or elsewhere could hide sensitive credentials. A helper binary on your computer would then control spawning new disposable VMs with premade images and your SSH key. The images can be lightweight or the desktop version with a remote VNC.
jboss10 15 hours ago [-]
Gondolin[1] is what you are describing. It's made by the same person who made the Pi coding agent and sends all of the agent's bash into a small QEMU vm.
He described a separate machine where the execution and context would be hosted, not local sandboxes. Did I misunderstand how Gondolin works?
binsquare 13 hours ago [-]
I agree with this, virtual machines are invented to solve the sandboxing/multi-tenant issues.
This is why ec2 and the likes all sell you access to virtual machines (dividing up their underlying hardware).
28304283409234 15 hours ago [-]
I use vagrant on a seperate machine in a seperate network. The magic of ssh makes it transparent for me, and I feel pretty sure the agents cannot get to stuff that matters.
ulrikrasmussen 11 hours ago [-]
Exactly the same setup for me. It works great and the whole setup consists of a single Vagrantfile which I maintain via Claude itself. All relevant dotfiles are synced to a folder on the host so e.g. conversations are preserved across reprovisioning of the VM. It's a simple solution, but very flexible.
Imustaskforhelp 14 hours ago [-]
quickemu[0] is sort of amazing for these type of use cases actually. I feel like I could use it more for this type of stuff.
I was using VMs for agents, but wanted something lighter and faster, so I made flar, which bubblewraps the agent, its config/history and the project directory. The agent runs in your usual environment but has no access to anything other than what it's been explicitly granted, short-circuiting prompt injections (or intentional secret exfiltration on the part of the agent or model) as well as any supply chain exploits the agent might accidentally introduce.
It starts instantly, as it's a namespace, rather than a full VM or container that has to be downloaded/built/updated on start.
It defaults to dangerously skip permissions mode, but is much safer than the very porous sandbox the agents provide, and the agent can't reach outside of it even if told to, by the user or a prompt injection.
einhard 16 hours ago [-]
I am certainly no expert in this space so it is quite possible I'm missing something critical, but what seems to work for me is a Podman image I built on my computer with some basic things I need (using OpenCode, but I imagine any other agent could be used instead):
From there I just run the Podman image from the command line (using a Fish function) that mounts the specific project I'm working on to /workspace. I guess there might be some vulnerabilities with shared kernels and such, but it seems like an easy way to have some isolation.
qznc 13 hours ago [-]
As opencode is inside your container, credentials and API keys are also inside the container. Prompt injection when your agent fetches some web site could have your agent leak this credentials to someone.
Also, do you restrict networking or does your container have full access to your internal network?
einhard 11 hours ago [-]
I restrict networking via my pfSense router, using VLANs. I also don't include any credentials in the container; the agents there can't push, pull, or really access anything else beyond the standard tools without any authentication.
I don't use Claude or any other paid agent at the moment, so if that were to change I'd probably modify the way I run this, but with this simple set up I'm not too worried about credentials leaking.
petesergeant 15 hours ago [-]
This is how I started, and then I wanted to bring along creature comforts, not have to re-auth per box (for subscription models), had some skills I usually wanted to bring, wanted slightly different setups for different stacks, sensibly install multiple agents, import git identity (but not credentials), mount other code folders ro but only for certain projects, etc etc, and ended up with a full Docker wrapper.
einhard 15 hours ago [-]
Ah, that makes sense. I've only recently started playing with this stuff, and I've been focusing a lot of just getting somewhat good at using LLMs for projects while developing my own intuition, so everything I do just uses the mounted project and uv directly. As soon as things get more complex, I imagine I'll end up with a full wrapper as well.
Unrelated: I enjoyed your latest blog entry. I recently starting thinking about how to show the work that is done with AI, and how we talk about it. I haven't come to any major conclusions (I wish!), but your post about the prompting being distinct from the actual work resonates with me. Reminds me somewhat of discussions about the art of photography compared to the art of editing photos as a distinct skill.
matheusmoreira 16 hours ago [-]
> clawk forward add my-project 3000
> clawk network allow my-project api.example.com
Can you describe the implementation details? How did you implement the firewall without root?
I vibecoded virtdev, a virtual machine orchestration project just like this one:
It was designed to not require root, and the nftables firewall ended up becoming the only exception. I'm very curious about how you implemented this. Did you find a better way?
westurner 16 hours ago [-]
(Years ago I had puppet and cobbler provisioning VMs over PXE and then iPXE. FWIU foreman is more actively maintained than cobbler, which is built on Django web framework.)
Vagrant manages VMs and virtual networks, in Ruby.
ansible-molecule creates, converges, and destroys VM(s) and containers, in order to test ansible playbooks and ansible roles in clean build roots.
`podman kube play` over `podman machine` might solve for agents that need multiple VMs/containers
- Podman Desktop can work with the same local k8s setups as Docker Desktop. Though there's certainly more state to manage with k8s for agent session farm, k8s probably has better logging and quotas than a VM management script on each node.
OpenShift on OpenStack is one way to do containers over VMs over bare metal.
Microshift also does container-selinux.
There is not an apparmor policy set for containers?
bwrap and liboverlayfs and libseccomp are almost but not quite containers.
There are stronger container isolation layers that are more like a full or lightweight VM, that might be better for agent sessions: gVisor, firecracker vm, Todo
Cloudflare workerd is the open source part of cloudflare workers, which run lightweight WASM and JS VMs with multi-tenant isolation.
It takes far less resources to run a cloudflare worker than to run a container on cloudflare. So, if it's possible for an agent to operate within a WASM runtime ~container, that's probably more optimal for agent sessions.
Cloudflare/artifact-fs does lazy shallow git clones with a FUSE filesystem.
docker and podman support multiple WASM runtimes for running WASM containers, e.g. for agent sessiobs
celrenheit 16 hours ago [-]
Thanks! There's no packet firewall at all, no iptables/nftables. On macOS the VM's NIC is a Virtualization.framework file-handle device. The daemon runs gvproxy, which terminates the guest's connections and re-dials them as host sockets, so I filter with an allow-list right before the dial.
One caveat, since you asked about root specifically: that's the macOS path, and it only works thanks to the fd NIC. Firecracker on Linux only speaks a TAP, which needs root, so there I do shell out to sudo, but just for the device. The filtering is still the same userspace allow-list.
matheusmoreira 15 hours ago [-]
Thanks for the pointers!! I'm using passt, didn't know about gvproxy. This is awesome and could provide rootless filtering and firewall for my guest VMs. I'm going to experiment with this!
celrenheit 15 hours ago [-]
My pleasure! You can find my fork here: https://github.com/clawkwork/gvisor-tap-vsock — the diff is tiny, just a hook in the TCP/UDP/ICMP forwarders that consults an allow-list before dialing
kstenerud 16 hours ago [-]
yoloAI does something similar:
- Sandbox on Linux using Docker, Podman, containerd, gVisor, Kata, Firecracker
- Sandbox on Mac using Docker (Docker Desktop or Orbstack), Podman, Apple containers, Seatbelt, Tart (Tart lets you run simulators).
- Network restriction
- Secrets control (file mounts or credentials broker)
- NO ambient data (ENV is replaced with a minimal and local-to-sandbox one, no host-side filesystem access beyond what you explicitly allow)
- Workdir protection: Your work dir is never modified until you apply the changes, either standalone or as a git commit. You can also diff before applying. Git runs SANDBOX side in case the repo has filters.
- Uses copy-on-write if your filesystem supports it (most modern ones do)
- Has built-in support for claude, codex, gemini, aider, and opencode, but you can also launch it in "shell" mode and run whatever you want.
- Supports VS code tunnels, so you can remotely access in VS code if you don't want to use the terminal.
- Layered API (golang) if you want to sandbox other things
- Self-contained binary. No external requirements other than the backends you want to use. Defaults to a ~/.yoloai dir for config/data, but you can point it anywhere.
Hadn't seen yoloai before. I really like new/diff/apply/destroy workflow, that's interesting. For my own needs the two things I was after were multi-repo worktrees (one sandbox spanning several repos, each on with its own worktree) and a single network restricted VM path I fully control, rather than many backends (I started with many backends at first but it was awkward to add network filtering to them).
Lots of overlap though, nice work, and I'll be reading through yours.
LuD1161 11 hours ago [-]
With agentjail ( https://github.com/LuD1161/agentjail ), I've tried to contain coding agents in os-native sandboxes (sbpl for macos and similarly for linux, <4ms start time) + policy guardrails evaluated by Open Policy Agent (OPA), policies written in rego.
Protocol aware network proxy coming soon
Then you can match a DSL and block particular network requests.
This ensures you no longer fear --dangerously-skip-permissions and stop babysitting agents
What else would you want to see in this project? Please star the repo, if you like the idea :)
yencabulator 8 hours ago [-]
> What else would you want to see in this project?
Absence of spamming mentions of it everywhere.
LuD1161 46 minutes ago [-]
You mean, I should stop posting much about it ? Is it too aggressive posting
Am sorry about that.
janalsncm 11 hours ago [-]
It seems like your approach depends on figuring out what the command is doing, classifying it into three tiers (deny/ask/approve) and then letting the agent run its command depending on that classification. Is that right?
LuD1161 35 minutes ago [-]
[dead]
pshirshov 11 hours ago [-]
[flagged]
rock_artist 16 hours ago [-]
> Requires macOS 14+ on Apple silicon. (Linux is supported via firecracker and currently experimental...)
So should be noted it's mostly macOS out of the box with some Linux support if I understand right.
celrenheit 15 hours ago [-]
mostly macOS out of the box. I've tested it on linux and it worked but it's not my daily driver. I'll make the platform scope clearer in the README
skybrian 16 hours ago [-]
If the agent is running on your machine, it will suspend when you put your laptop asleep. I prefer using a remote Linux VM to let the coding agent keep working.
I’m quite happy with exe.dev for this. My laptop is asleep upstairs but I have an agent coding away in a browser tab on the tablet I’m using. I could also check on it from my phone.
But it might also be nice if a setup similar to exe.dev were available for self-hosting. I have a Mac Mini that I don’t really use much.
nzoschke 15 hours ago [-]
Another +1 to cloud sandboxes (and exe.dev) vs laptop sandbox.
Runs 24/7 completely air gapped from your laptop.
You also want a service proxy so the sandbox can access GitHub or Stripe without keys accessible to the agent. I haven't seen many of the laptop sandbox tools do this, where exe.dev does it out of the box with their "integrations".
I usually drop my own binary agent coding toolkit inside the sandbox so I have things like a code browsing and review right there in every sandbox too.
Like you I also have a Mac Mini I've thought about making into my own 24/7 dev box, but building this vs buying 50 VMs from exe.dev for $20/mo doesn't add up for me.
matheusmoreira 16 hours ago [-]
> it will suspend when you put your laptop asleep
It is possible to simply not do that. Laptops work just fine as servers. They even have a builtin monitor and UPS.
skybrian 15 hours ago [-]
Yeah but I ‘d rather run it remotely and not run agents on my laptop at all.
mswphd 16 hours ago [-]
https://paseo.sh/ supports self-hosting, though I've only used it a mild amount tbh.
skybrian 15 hours ago [-]
It looks like they have the right idea but their workspaces are based on GitHub worktrees rather than separate VM’s.
celrenheit 14 hours ago [-]
[dead]
arjie 14 hours ago [-]
I just use kubernetes with a single PVC for the Claude folders. Works like a charm. Don’t need too much else. You can then start separate pods with different PVCs to different folders. Lives on a server, I can manage the auth by just making each pod description different.
Honestly works fine and you can use /rc to talk to them over the Claude phone app.
- Drives Claude Code as a quasi-subagent via "Channels," which supports multi-turn interaction with the host and suspend / resume
- Declarative configuration in the flake of exactly what is copied into the VM (besides the local project), what DNS origins are allowed, etc.
- Shared access to host Nix store / object DB via overlay FS
- Syncs code with the host over a shared (local) Git remote (no worktree mess to manage)
- Devshell commands for quickly dropping into the guest and viewing status of all VMs etc.
- VMs start in a couple of seconds
riedel 13 hours ago [-]
Nice I started to vibecode something with bwrap.nix: https://github.com/riedel/nix-opencode-with-mitm (not anywhere near production ready) What I would really want is token injection outside the sandbox and good control over the network, why I tried integrating a mitm proxy outside the sandbox.
mkagenius 13 hours ago [-]
If you are into nixos, we at InstaVM have just launched[1] nixos based sandboxes.
That's very cool, although AFAICT my hacked-together non-product shares most of the same virtues. Can you upsell it to me a little?
mkagenius 12 hours ago [-]
Our nixos sandboxes are mostly a cloud offering at this time. Scale is what we can offer additionally. Have secret injection built in too. I see you have suspend, resume we have the same. We have snapshots and cloning. Startup time is sub second with warm pools - which will go further down when we replace firecracker with Tarit soon.
So many of these agent sandboxing solutions now. Don’t people search for existing solutions anymore?
lschueller 8 hours ago [-]
I'm afraid not. My impression is, most are just prompting for the tool they need right now, accept bugs and maybe even security gaps, but save time and pain not searching for an established tool. Which is fine and understandable for private things. But I wonder, how often this happens in professional environments. But maybe I'm just wrong... I have not seen any resarch or evaluation for this claims. Would be interesting, though
djha-skin 15 hours ago [-]
Can't I just use a `docker run` command to run claude and bind in the current directory into the docker container?
A lot of people would yell "docker isn't secure!" but if you're running Docker on MacOS (and Windows, I think) you're already running a Linux VM. But if you really wanna let your agent YOLO their way along, Docker is kinda restrictive, so that a full VM lets it install things, start services, and so on.
Which is just a front for systemd-nspawn. It's annoying you have to edit the config.nspawn to mount a directory if you start it with the "shell" command, instead of booting. But apart from that, it's brilliant.
When firing up agent VMs for my personal projects, I tried to completely separate SSH keys as well. Surprisingly annoying to have multiple keys for, say GitHub, and restrict the agent to use a deploy key.
I saw this project enables ssh-agent forwarding, so my question; is this a non-issue to begin with? Or just not your focus currently.
It is docker + vscode friendly.
I tested it with major systems (copilot, codex, Claude Code and pi.dev)
Comments Wellcome!
vqtska 16 hours ago [-]
I still don't understand the point of all these VMs and containers for agents. Just create a separate user on your machine without sudo privileges, switch to it in your terminal and run all the agents you want without it being able to reach your files. What am I missing?
killerstorm 16 hours ago [-]
Privilege escalation (e.g. setuid), world-readable files might contain sensitive data, world-writeable files, unrestricted network access (including access to all locally running services)... If you have fully patched system without zero-days and it's configured in a perfect way, then, sure...
Container is quite like a "separate user" except you can explicitly define what it can access.
(Even if all your daemons have good auth, it's now quite common for _apps_ to open listening sockets without much auth...)
wilkystyle 16 hours ago [-]
Also, many of these sandboxing solutions provide features like network allowlists and credential masking/injection
vqtska 16 hours ago [-]
Sure, if you assume the agent will be hostile on you. I thought it's just so the agent doesn't accidentally rm -rf / on you
killerstorm 16 hours ago [-]
The agent might install hostile software, e.g. a npm package. Unfortunately, very common problem nowadays.
And that's without anything like prompt injection happening.
pigeons 16 hours ago [-]
They do try privilege escalation unprompted.
matheusmoreira 16 hours ago [-]
You're missing the fact you'd be sharing a kernel with the sandboxed agent. Virtualization presents an infinitely smaller attack surface.
zzril 16 hours ago [-]
If your threat model is that of a malicious agent that will use a 0-day LPE to get root and exfiltrate all of your SSH keys, virtualization makes sense. But then, I wouldn't run such an agent at all, if not specifically in the context of malware analysis.
If you're just concerned about "agent messing up and taking the rules in some markdown files more laxly than I would have", then running it as a seperate user is totally enough...
matheusmoreira 16 hours ago [-]
Threat model is supply chain attacks on unmaintained package repositories like npm, pip and cargo. Everything on my host comes from my Linux distribution's repositories. Everything else gets virtualized with extreme prejudice. I'll even virtualize my Steam games one of these days.
khuey 16 hours ago [-]
IMO the threat model is more letting the agent loose on issues/PRs in a public Github repo and it getting tricked into running a malicious payload.
vqtska 16 hours ago [-]
What kind of things are you even doing that the agent would try to perform a kernel exploit on you? I thought sandboxing is just to protect from the agent accidentally clearing your home directory.
Side note, just 6 days ago a Linux VM escape exploit was disclosed.
matheusmoreira 16 hours ago [-]
I'm not worried about the agent at all. The VM is there to prevent it from clobbering files on my real system.
I'm worried about supply chain attacks on npm, pip, cargo and everything else. Don't want to get compromised if I install some stupid package.
My virtdev project has essentially split my computer into two systems: my "real" trusted system with software coming directly from my Linux distribition's repositories, and the VMs for everything else.
> just 6 days ago a Linux VM escape exploit was disclosed
Well, shit. Details?
SoftTalker 15 hours ago [-]
> npm, pip, cargo and everything else
All that stuff should also go into the agent user's home directory.
Patched in Linux 7.1.3, which happens to be the exact kernel version I'm running. Update your kernels, folks.
bheadmaster 16 hours ago [-]
VMs and containers are fairly isolated and reproducible. A separate user on your machine still depends on the programs installed locally.
Greenpants 16 hours ago [-]
And those installed programs could have vulnerabilities that just a non-root user account could still take advantage of. Perhaps not likely for an LLM to do, but more so if you let them loose on the internet and they end up coming across prompt injection that instructs exactly that :)
3form 16 hours ago [-]
Well, for one, Debian and Debian-based distros make your home directory readable by everyone by default.
Security is riddled by traps. If you can afford best possible level of isolation, why not do it?
pmontra 13 hours ago [-]
Yep, I broke locate when I made my home 700. Its user could not traverse my files anymore. I had to make it run as root. A better design would be to traverse each user home with that user id but apparently it assumes that the home dirs are 755.
Codex uses bwrap sandbox which is purely cosmetic - out-of-box it sees your ssh keys and can send it to remote server. It prevents agent from deleting outside files by mistake, but does nothing against malicious activity
bpavuk 16 hours ago [-]
that Pi extension, by contrast, gives you a popup whenever an agent tries to access something outside the current working directory and send a network request. that depends on configuration. also if agent tries to ssh into something that will also send a request for access, so child processes are covered
0xbadcafebee 16 hours ago [-]
Actually that doesn't just work on all systems and it breaks on others. The alternate user is actually guaranteed to work on all systems and it's built in
bpavuk 16 hours ago [-]
Seatbelt on macOS and Landlock over WSL2 on Windows 11. Landlock is also built-in, and WSL2 is a first-party downloadable component
celrenheit 15 hours ago [-]
There are two sides to this. The first is security, which plenty of comments already covered. The second, and the real one for me: my tests spin up Docker containers, and I was building a Kubernetes tool (argocd/flux style) that needs a real cluster in the sandbox. In a container that means Docker-in-Docker, which always felt hacky. A VM is just a normal Linux box where Docker and k8s run like they do everywhere else. A separate user can't give you that, it shares your one kernel and whatever's already installed on the host.
schainks 15 hours ago [-]
There is a reason why VMs even are a thing at all: they can offer better security guarantees than alternatives.
A separate user is a good start but LLM tests themselves show they can cleverly bypass guardrails if they figure out they are in a sandboxed environment of some kind, right?
So, I read those test results as: an LLM is less likely to do something crazy if it thinks it has the whole environment to itself.
de6u99er 16 hours ago [-]
On my Mac, every sudo requires either my fingerprint or password, and times out immediately.
qznc 15 hours ago [-]
Colleague recently told me his agent tried to sudo. It failed but it saw that docker with root is available and just used that. I don't quite remember what it did, I think editing something in /etc.
throw1234567891 16 hours ago [-]
And why would I do that when I can create a vm from a snapshot and be done in 30 seconds?
augusto-moura 16 hours ago [-]
Mainly networking and namespaces, same reasons why we run services on docker instead of old multi-user setups
altcognito 16 hours ago [-]
In a corporate environment they may not have that option
SoftTalker 16 hours ago [-]
But they can run arbitrary VMs?
petesergeant 16 hours ago [-]
am I doing that once per project, or?
TheRealPomax 16 hours ago [-]
If you think those two contexts are equally secure, your machine is incredibly insecure and I hope you run daily incremental full system backups.
You have far too much data in unsecured locations, and you have far too little understanding of what an agent would do, to go "I trust whatever this user account will be doing on my machine".
sxiong 12 hours ago [-]
We need an agentic platform that has sandboxing built-in well. Giving agents a disposable VM is nice but I'm already hitting my Mac mini's space limit.
mcintyre1994 11 hours ago [-]
Fly.io’s sprites.dev is a similar idea but on their hardware and you get a public/private domain to run a server too, I’ve found it pretty good.
giancarlostoro 12 hours ago [-]
Claude Cowork does this on Mac and ears several gigabytes of on disk storage, drives me crazy, because it does this if you click on it without ever using it.
zhshhan 10 hours ago [-]
[flagged]
bad_haircut72 16 hours ago [-]
Sprites from Fly io does this beautifully, claude comes preinstalled, its great
whitlock 15 hours ago [-]
The core abstraction seems identical to Fly.io Sprites: give the coding agent its own real Linux machine, put a hard network/isolation boundary around it, then yeet.
Sprites arguably has the better security boundary, since the agent isn’t sitting adjacent to my laptop and home network.
This is like the 30th AI sandbox project on Show HN. Why this one over the rest?
SoftTalker 16 hours ago [-]
Also the readme (and thus likely the whole thing) is clearly LLM-generated.
15 hours ago [-]
celrenheit 15 hours ago [-]
[dead]
aftbit 16 hours ago [-]
Does Codex run its own sandbox? I see that sometimes it runs commands without asking, which then fail for some reason, and it asks to run them again "outside of the sandbox"
Cursor has something similar. I don't know about Claude Code but I assume it does as well since Anthropic has open sourced their own sandboxing tool.
ttul 16 hours ago [-]
And, make coding harnesses run 20x faster at many multi-processing and file operations by running the coding harness itself inside of Linux instead of MacOS…
celrenheit 14 hours ago [-]
That's actually how clawk runs it: the agent process (claude/codex) runs inside the Linux guest on a PTY, not on macOS
mlnj 16 hours ago [-]
Is file IO a bottleneck now?
ttul 12 hours ago [-]
A lot of the work that coding agents do requires tons of random access to files, which MacOS is particularly slow at. Not because the hardware is bad, mind you. Mac SSDs are amazing. It’s the file system. Secondly, MacOS supervises processes for security reasons and coding harnesses spawn tons of processes. There is a lot of CPU overhead embedded in that supervision.
polotics 12 hours ago [-]
Ok! why should I switch away from running Firejail on Fedora? Ok SUID security thanks
luciana1u 12 hours ago [-]
the natural evolution of AI tooling: first we gave them our laptops, then we gave them VMs, next they'll demand their own AWS accounts and a corporate card
grzes 14 hours ago [-]
i was quite shocked to learn that cursor is not scoped to current workspace, recently it just traversed my projects dir outside my current workspace and looked for patterns in code.
anoop_kumar 16 hours ago [-]
why not just use something like smol VMs? So many VM's and I am confused on which one to use for what. I think we now need a VM orchestrator!
windex 15 hours ago [-]
I use a raspberry pi for all my experiments
JeremyNT 13 hours ago [-]
This is the right idea IMO but I don't want to trust a third party tool for it either (no offense OP!).
Ever since I started running coding agents with shell access I've jailed them inside of VMs. Since I run Linux I can just use incus[0] for the VM management layer.
It's extremely simple, and you can vibe code a shell script to customize your workflow in a few minutes.
Because that means you are sharing kernel with the sandboxed agent. Virtualization presents an infinitely smaller attack surface.
hwc 16 hours ago [-]
If there is any attack surface within a properly-configured container, that's a kernel bug, right?
matheusmoreira 16 hours ago [-]
Probably. If I remember correctly, containers on Linux are implemented using the kernel's namespaces. The same ones which became famous for the vulnerabilities they surfaced in previously unexercised code.
croes 16 hours ago [-]
Then use SmolVM
matheusmoreira 15 hours ago [-]
I'm already using virtdev, my own solution built on top of QEMU.
0xbadcafebee 16 hours ago [-]
How many of these are there now, a hundred? We get it, you can run an agent in a VM/container/sandbox. What about configuration management & rollbacks? What about the policy engine? What about dynamic credential management? What about the lethal trifecta? A sandbox is the easiest part and doesn't address the others.
matheusmoreira 15 hours ago [-]
Huh. My virtdev project implements nearly all of that... Except credential injection from the host, which turned out to be on my TODO list.
0xbadcafebee 12 hours ago [-]
Based on the readme in your project, yours does not include 1) a policy engine, 2) application proxies to limit scope of access to external systems/networks, 3) firewall configuration for commonly accessed package repositories, 4) configuration management, 5) credential management.
You do have some scripts that 'diff' package inventory of the one distro you support, but not a full fledged configuration management system to manage package dependencies, file permissions, services, etc. Technically the user can deploy one using your provisioning hook but since it's not built-in it's another component the user will need to bring with them, which is one more reason they can use any other system that does basic sandboxing. I'm sure it's useful for you, but it doesn't do anything all the other solutions don't already do. You basically made Vagrant but without the useful Vagrantfile and OS-agnosticism.
matheusmoreira 10 hours ago [-]
You quietly dropped rollbacks and the lethal trifecta from your list, the two exact things my project is great at. It's got qcow2 delta images and a fully customizable egress firewall backed by standard nftables.
You substituted in application proxies and firewalls for package repositories. Implementing those is what I came to this thread for. Already planning a custom network stack to replace passt. It will have those features soon. Credential management too.
> It doesn't do anything all the other solutions don't already do
That's just false. I built virtdev because I literally didn't find any other tool that implemented KVM virtualization, cheap expendable contextual VMs and configurable egress firewall with minimal, soon to be zero root access requirements. Virtdev also manages daemon life cycle correctly via user mode systemd, which is something I just don't see other projects do.
Vagrantfile and OS agnosticism are not why I built virtdev. Vagrant has no security focus at all, and I explicitly opted out of declarative YAML because GitHub Actions is painful enough.
> since it's not built-in
By this logic, no composable tool has any value. I chose to provide mechanism, not policy. The primitives are there.
> it's another component the user will need to bring with them
Yes, as files committed to a dotfiles repository. A one time configuration.
selvakn 12 hours ago [-]
shameless plug (im a contributor to):
https://github.com/jskswamy/aide/ covers you, with configuration driven through yaml and credential management with sops.
oleg_kabanov 15 hours ago [-]
What does Clawk do that Docker without exposed ports doesn't?
ebeirne 16 hours ago [-]
love what youve done here. i will be using this in the future.
celrenheit 16 hours ago [-]
Thank you, that means a lot !
htrp 16 hours ago [-]
how does this compare with the aws lambda microvms?
TheRealPomax 16 hours ago [-]
> On your own machine that leaves two bad options. You approve every command (and babysit a prompt every few seconds), or you run --dangerously-skip-permissions and hope nothing important is one rm -rf or one leaked token away.
Literally everyone has the option to use a VM - it's built into Windows, UTM on MacOS, Docker on Linux. Yes, "a tool that automatically builds a VM" is useful, but we've had a third option (four, if you count "actually I disagree with the idea that it's only useful if it's fully agentic") from day one.
celrenheit 15 hours ago [-]
[dead]
mlnj 16 hours ago [-]
Is there anyone who is NOT building another AI sandbox solution?
You might prefer byre's simplicity, transparency, and ease of reasoning about: one local container, explicit access grants, readable generated Docker, and a workflow that stays close to normal development rather than introducing a larger sandbox platform. It's also very very easy to eject from if you want to stop using it.
elombn 16 hours ago [-]
[flagged]
sumar7 17 hours ago [-]
[flagged]
ronalder100 12 hours ago [-]
[flagged]
linggen 13 hours ago [-]
[dead]
escape_42 16 hours ago [-]
[dead]
bitwize 16 hours ago [-]
Errbody gangsta until the agent figures out it's in a container and finds an exploit that lets it break out of container jail...
matheusmoreira 16 hours ago [-]
I developed a VM project just like this one. Asked Fable to stress test it and try to break out of containment, and to my surprise it didn't manage to. Fable didn't get downgraded to Opus either, for some reason.
Would have thrown Mythos at it if I had access to it.
arcanemachiner 16 hours ago [-]
Isn't Fable just Mythos + guardrails? Sounds like you did throw Mythos at it.
matheusmoreira 16 hours ago [-]
I initially thought that was the case: literally the same weights but with an incredibly obnoxious "safety classifier" tacked on. Now I'm uncertain because people on HN have told me it's a different model altogether with further fine tuning for safety or something.
nullc 16 hours ago [-]
You should conduct the same test with knowingly faulty containment, otherwise the theory that the model is hobbled should probably outrank that it couldn't escape.
matheusmoreira 16 hours ago [-]
That's a good idea. I don't have Fable at the moment, when it comes back I'll ochestrate those tests.
Opus did manage to iron out a lot of fail open bugs during development though, and Fable's like a hundred times more relentless than Opus. I'm not saying it's a definitive result or that my VM firewall thingy is unhackable... I'm just saying it put a smile on my face.
16 hours ago [-]
mrbn100ful 16 hours ago [-]
YAY, let's build the same thing over and over :/
Did you know that before reinventing the wheel, you can ask LLM your requirements, and they can find the most suited already existing tool ?
matheusmoreira 15 hours ago [-]
Yeah, but I explicitly chose to make my own anyway because I wanted to fully own the code and none of the existing solutions worked the way I wanted.
There is absolutely nothing wrong with reinventing the whell. It's entirely possible to do it and end up discovering you've made a better wheel.
28304283409234 15 hours ago [-]
`vagrant snapshot` exists.
croes 16 hours ago [-]
Isn’t the point of LLM to recreate the wheel?
No libraries, no frameworks. Let AI recreate it from scratch … every time
[1](https://earendil-works.github.io/gondolin/
This is why ec2 and the likes all sell you access to virtual machines (dividing up their underlying hardware).
[0]: https://github.com/quickemu-project/quickemu
https://github.com/swelljoe/flar
It starts instantly, as it's a namespace, rather than a full VM or container that has to be downloaded/built/updated on start.
It defaults to dangerously skip permissions mode, but is much safer than the very porous sandbox the agents provide, and the agent can't reach outside of it even if told to, by the user or a prompt injection.
Also, do you restrict networking or does your container have full access to your internal network?
I don't use Claude or any other paid agent at the moment, so if that were to change I'd probably modify the way I run this, but with this simple set up I'm not too worried about credentials leaking.
Unrelated: I enjoyed your latest blog entry. I recently starting thinking about how to show the work that is done with AI, and how we talk about it. I haven't come to any major conclusions (I wish!), but your post about the prompting being distinct from the actual work resonates with me. Reminds me somewhat of discussions about the art of photography compared to the art of editing photos as a distinct skill.
> clawk network allow my-project api.example.com
Can you describe the implementation details? How did you implement the firewall without root?
I vibecoded virtdev, a virtual machine orchestration project just like this one:
https://github.com/matheusmoreira/virtdev
It was designed to not require root, and the nftables firewall ended up becoming the only exception. I'm very curious about how you implemented this. Did you find a better way?
Vagrant manages VMs and virtual networks, in Ruby.
ansible-molecule creates, converges, and destroys VM(s) and containers, in order to test ansible playbooks and ansible roles in clean build roots.
podman machine manages VMs:
- podman-container-tools/podman-machine-os: machine image files: https://github.com/podman-container-tools/podman-machine-os/...
`podman kube play` over `podman machine` might solve for agents that need multiple VMs/containers
- Podman Desktop can work with the same local k8s setups as Docker Desktop. Though there's certainly more state to manage with k8s for agent session farm, k8s probably has better logging and quotas than a VM management script on each node.
OpenShift on OpenStack is one way to do containers over VMs over bare metal.
Microshift also does container-selinux.
There is not an apparmor policy set for containers?
bwrap and liboverlayfs and libseccomp are almost but not quite containers.
There are stronger container isolation layers that are more like a full or lightweight VM, that might be better for agent sessions: gVisor, firecracker vm, Todo
Cloudflare workerd is the open source part of cloudflare workers, which run lightweight WASM and JS VMs with multi-tenant isolation.
It takes far less resources to run a cloudflare worker than to run a container on cloudflare. So, if it's possible for an agent to operate within a WASM runtime ~container, that's probably more optimal for agent sessions.
Cloudflare/artifact-fs does lazy shallow git clones with a FUSE filesystem.
- "Show HN: VM-curator – a TUI alternative to libvirt and virt-manager" https://news.ycombinator.com/item?id=46750437
https://news.ycombinator.com/item?id=46825026 ; amla sandbox, agentvm, ARM64 MTE
https://news.ycombinator.com/item?id=46825119 ; container2wasm , vscode-container-wasm-gcc-example ; build WASM containers with Dockerfiles
docker and podman support multiple WASM runtimes for running WASM containers, e.g. for agent sessiobs
- Sandbox on Linux using Docker, Podman, containerd, gVisor, Kata, Firecracker
- Sandbox on Mac using Docker (Docker Desktop or Orbstack), Podman, Apple containers, Seatbelt, Tart (Tart lets you run simulators).
- Network restriction
- Secrets control (file mounts or credentials broker)
- NO ambient data (ENV is replaced with a minimal and local-to-sandbox one, no host-side filesystem access beyond what you explicitly allow)
- Workdir protection: Your work dir is never modified until you apply the changes, either standalone or as a git commit. You can also diff before applying. Git runs SANDBOX side in case the repo has filters.
- Uses copy-on-write if your filesystem supports it (most modern ones do)
- Has built-in support for claude, codex, gemini, aider, and opencode, but you can also launch it in "shell" mode and run whatever you want.
- Supports VS code tunnels, so you can remotely access in VS code if you don't want to use the terminal.
- Full lifecycle support: Launch, attach, stop, restart, wait, one-shot, clone, destroy
- MCP passthrough
- Layered API (golang) if you want to sandbox other things
- Self-contained binary. No external requirements other than the backends you want to use. Defaults to a ~/.yoloai dir for config/data, but you can point it anywhere.
- FOSS
https://github.com/kstenerud/yoloai
Protocol aware network proxy coming soon Then you can match a DSL and block particular network requests.
This ensures you no longer fear --dangerously-skip-permissions and stop babysitting agents
What else would you want to see in this project? Please star the repo, if you like the idea :)
Absence of spamming mentions of it everywhere.
So should be noted it's mostly macOS out of the box with some Linux support if I understand right.
I’m quite happy with exe.dev for this. My laptop is asleep upstairs but I have an agent coding away in a browser tab on the tablet I’m using. I could also check on it from my phone.
But it might also be nice if a setup similar to exe.dev were available for self-hosting. I have a Mac Mini that I don’t really use much.
Runs 24/7 completely air gapped from your laptop.
You also want a service proxy so the sandbox can access GitHub or Stripe without keys accessible to the agent. I haven't seen many of the laptop sandbox tools do this, where exe.dev does it out of the box with their "integrations".
I usually drop my own binary agent coding toolkit inside the sandbox so I have things like a code browsing and review right there in every sandbox too.
https://github.com/housecat-inc/scratch
Like you I also have a Mac Mini I've thought about making into my own 24/7 dev box, but building this vs buying 50 VMs from exe.dev for $20/mo doesn't add up for me.
It is possible to simply not do that. Laptops work just fine as servers. They even have a builtin monitor and UPS.
Honestly works fine and you can use /rc to talk to them over the Claude phone app.
Some highlights:
- Drives Claude Code as a quasi-subagent via "Channels," which supports multi-turn interaction with the host and suspend / resume
- Declarative configuration in the flake of exactly what is copied into the VM (besides the local project), what DNS origins are allowed, etc.
- Shared access to host Nix store / object DB via overlay FS
- Syncs code with the host over a shared (local) Git remote (no worktree mess to manage)
- Devshell commands for quickly dropping into the guest and viewing status of all VMs etc.
- VMs start in a couple of seconds
1. https://instavm.io/blog/nixos-reproducible-dev-environments-...
1. Tarit - https://GitHub.com/instavm/tarit
Seems like both projects are following very similar approaches.
A lot of people would yell "docker isn't secure!" but if you're running Docker on MacOS (and Windows, I think) you're already running a Linux VM. But if you really wanna let your agent YOLO their way along, Docker is kinda restrictive, so that a full VM lets it install things, start services, and so on.
Which is just a front for systemd-nspawn. It's annoying you have to edit the config.nspawn to mount a directory if you start it with the "shell" command, instead of booting. But apart from that, it's brilliant.
I saw this project enables ssh-agent forwarding, so my question; is this a non-issue to begin with? Or just not your focus currently.
https://bitrise.io/platform/remote-dev-environments
Bitrise has been doing macOS VMs for CI for 10 years, so we extended the existing product to this use-case.
https://github.com/daitangio/take-ai-control
It is docker + vscode friendly. I tested it with major systems (copilot, codex, Claude Code and pi.dev) Comments Wellcome!
Container is quite like a "separate user" except you can explicitly define what it can access.
(Even if all your daemons have good auth, it's now quite common for _apps_ to open listening sockets without much auth...)
And that's without anything like prompt injection happening.
If you're just concerned about "agent messing up and taking the rules in some markdown files more laxly than I would have", then running it as a seperate user is totally enough...
Side note, just 6 days ago a Linux VM escape exploit was disclosed.
I'm worried about supply chain attacks on npm, pip, cargo and everything else. Don't want to get compromised if I install some stupid package.
My virtdev project has essentially split my computer into two systems: my "real" trusted system with software coming directly from my Linux distribition's repositories, and the VMs for everything else.
> just 6 days ago a Linux VM escape exploit was disclosed
Well, shit. Details?
All that stuff should also go into the agent user's home directory.
CVE-2026-43499 - https://nebusec.ai/research/ionstack-part-2/
Security is riddled by traps. If you can afford best possible level of isolation, why not do it?
A separate user is a good start but LLM tests themselves show they can cleverly bypass guardrails if they figure out they are in a sandboxed environment of some kind, right?
So, I read those test results as: an LLM is less likely to do something crazy if it thinks it has the whole environment to itself.
You have far too much data in unsecured locations, and you have far too little understanding of what an agent would do, to go "I trust whatever this user account will be doing on my machine".
Sprites arguably has the better security boundary, since the agent isn’t sitting adjacent to my laptop and home network.
Cursor has something similar. I don't know about Claude Code but I assume it does as well since Anthropic has open sourced their own sandboxing tool.
Ever since I started running coding agents with shell access I've jailed them inside of VMs. Since I run Linux I can just use incus[0] for the VM management layer.
It's extremely simple, and you can vibe code a shell script to customize your workflow in a few minutes.
[0] https://linuxcontainers.org/incus/
You do have some scripts that 'diff' package inventory of the one distro you support, but not a full fledged configuration management system to manage package dependencies, file permissions, services, etc. Technically the user can deploy one using your provisioning hook but since it's not built-in it's another component the user will need to bring with them, which is one more reason they can use any other system that does basic sandboxing. I'm sure it's useful for you, but it doesn't do anything all the other solutions don't already do. You basically made Vagrant but without the useful Vagrantfile and OS-agnosticism.
You substituted in application proxies and firewalls for package repositories. Implementing those is what I came to this thread for. Already planning a custom network stack to replace passt. It will have those features soon. Credential management too.
> It doesn't do anything all the other solutions don't already do
That's just false. I built virtdev because I literally didn't find any other tool that implemented KVM virtualization, cheap expendable contextual VMs and configurable egress firewall with minimal, soon to be zero root access requirements. Virtdev also manages daemon life cycle correctly via user mode systemd, which is something I just don't see other projects do.
Vagrantfile and OS agnosticism are not why I built virtdev. Vagrant has no security focus at all, and I explicitly opted out of declarative YAML because GitHub Actions is painful enough.
> since it's not built-in
By this logic, no composable tool has any value. I chose to provide mechanism, not policy. The primitives are there.
> it's another component the user will need to bring with them
Yes, as files committed to a dotfiles repository. A one time configuration.
Literally everyone has the option to use a VM - it's built into Windows, UTM on MacOS, Docker on Linux. Yes, "a tool that automatically builds a VM" is useful, but we've had a third option (four, if you count "actually I disagree with the idea that it's only useful if it's fully agentic") from day one.
You might prefer byre's simplicity, transparency, and ease of reasoning about: one local container, explicit access grants, readable generated Docker, and a workflow that stays close to normal development rather than introducing a larger sandbox platform. It's also very very easy to eject from if you want to stop using it.
Would have thrown Mythos at it if I had access to it.
Opus did manage to iron out a lot of fail open bugs during development though, and Fable's like a hundred times more relentless than Opus. I'm not saying it's a definitive result or that my VM firewall thingy is unhackable... I'm just saying it put a smile on my face.
Did you know that before reinventing the wheel, you can ask LLM your requirements, and they can find the most suited already existing tool ?
There is absolutely nothing wrong with reinventing the whell. It's entirely possible to do it and end up discovering you've made a better wheel.
No libraries, no frameworks. Let AI recreate it from scratch … every time