[woodruffw]

For one, because Docker is not a sandbox, and containers are not a strong security boundary[1]. What you really need here is a strongly isolated VM, at which point you're playing cat-and-mouse games with your target: their new incentive is to detect your (extremely detectable) VM, and your job is to make the VM look as "normal" as possible without actually making it behave normally (because this would mean getting exploited). That kind of work has a long and frustrating tail, and it's not particularly fruitful (relative to the other things packaging ecosystems can do to improve package security).

> I know there is a researcher out there who has retrieved and installed every single pip package to do an analysis, which is a good start.

You're probably talking about Moyix, who did indeed downloaded every package on PyPI[2], and unintentionally executed a bunch of arbitrary code on his local machine in the process.

[1]: https://cloud.google.com/blog/products/gcp/exploring-contain...

[2]: https://moyix.blogspot.com/2022/09/someones-been-messing-wit...

[eigenvalue]

You make some good points. But it still seems to me that, if you used the best available sandboxed VMs for each platform (Windows Sandbox for Windows; FireJail for Linux; VirtualBox with no folder permissions for OSX-- I don't know if these are the best or even good, those were the ones I found from a bit a searching), that you could install and run these packages in an automated way (especially with some GPT3-type help to figure out how to explore and call the important functions) and look for the telltale signs in the network and file access behavior that they are malicious. Even if we grant that this is a long-tailed "cat and mouse" game, then so what? We won't get 100% security, especially against super sophisticated threat actors, but if you could catch 98% or whatever of the typical clumsy supply chain attacks, or super egregious stuff like that NPM package that deleted your whole disk if you were Russian, that would be an incredibly vast improvement over the current state of affairs. Why isn't that worth doing? Why isn't Google or Microsoft at least trying this?

[woodruffw]

It isn't worth doing because the equation you've supplied doesn't include the effect of catastrophic failure: dynamic analysis lowers the barrier for exploit to a single hypervisor or VM exploit. Catching 98% of spam packages that affect nobody is worth very little when the 2% you don't catch are the ones that do the real damage.

> Why isn't Google or Microsoft at least trying this?

They are: Google and Microsoft both spend (tens of) millions of dollars on hypervisor and VM isolation research each year. It's a huge field.

[com2kid]

> What you really need here is a strongly isolated VM,

Simplify, don't use a VM.

Create an isolated network, hook your sacrificial machine up to it, have it install the package. Remotely kill it (network controlled power switch if needed). The machine's hard drive should be hooked up through a network controlled switch of some type. After the sacrificial machine is powered down, reroute the HD so it is connected to a machine that does forensics.

Now you have a clear "before" and "after" situation setup for analysis.

The sacrificial machine's network activity can be monitored by way of whatever switch/router it uses to connect to the Internet.

[woodruffw]

This is a VM, but flakier and with more steps! It’s also eminently not sustainable on PyPI’s scale, which is the context we’re talking about. I’m

[com2kid]

Doesn't it solve VM sandbox escape problems though? Actual physical hardware isolation, along with an isolated network. Code can't detect it is running on a VM if there isn't a VM, and it sure can't escape the sandbox if there isn't a sandbox.

> It’s also eminently not sustainable on PyPI’s scale, which is the context we’re talking about.

I started my software engineering career in testing before VMs were a thing, so large, very large, scale test setups like the one I outlined were common place. I wrote about some of my experiences at https://meanderingthoughts.hashnode.dev/how-microsoft-tested... and the physical hardware setup my team was using to run (millions of!) tests was tiny compared to what other teams in Microsoft did at the time.

Network controlled power and peripherals were exactly how automation was done back in the day. Instead of VM images, you got a bunch of identical(ish) hardware and you wrote fresh images to hard drives to reset your state.

Are VMs more convenient? Sure, but my reply was in context of ensuring malware can't detect it is running in a VM!