objectified

I was thinking the exact same thing. It gets a lot more attention on Youtube, really not understanding what's up.

> The moment you run LLM generated code, any hallucinated methods will be instantly obvious: you’ll get an error. You can fix that yourself or you can feed the error back into the LLM and watch it correct itself.

But that's for methods. For libraries, the scenario is different, and possibly a lot more dangerous. For example, the LLM generates code that imports a library that does not exist. An attacker notices this too while running tests against the LLM. The attacker decides to create these libraries on the public package registry and injects malware. A developer may think: "oh, this newly generated code relies on an external library, I will just install it," and gets owned, possibly without even knowing for a long time (as is the case with many supply chain attacks).

And no, I'm not looking for a way to dismiss the technology, I use LLMs all the time myself. But what I do think is that we might need something like a layer in between the code generation and the user that will catch things like this (or something like Copilot might integrate safety measures against this sort of thing).

The problem there though, is that with PoCs like this, as an attacker you want to have a ping back to your system so that you know the attack has been successful (in this case they probably expected/hoped someone at Cursor to install the package, that's the usual objective in a dependency confusion attack). But what they could have done, is send a less sensitive thing like just the current working directory or current effective user, instead of the whole environment.

At first I was excited to see that a new tool would solve the Python "packaging" problem. But upon further reading, I realized that this was about _package management_, not so much about packaging a Python application that I've built.

Personally I haven't had many problems with package management in Python. While the ecosystem has some shortcomings (no namespaces!), pip generally works just fine for me.

What really annoys me about Python, is the fact that I cannot easily wrap my application in an executable and ship it somewhere. More often than not, I see git clones and virtualenv creation being done in production, often requiring more connectivity than needed on the target server, and dev dependencies being present on the OS. All in all, that's a horrible idea from a security viewpoint. Until that problem is fixed, I'll prefer different languages for anything that requires some sort of end user/production deployment.

I'm not sure I understand your question correctly, but I'm talking specifically about outbound network traffic. Your API's application servers (where such evil libraries could be deployed) should not be able to have any network connectivity towards the internet. So on that server, you should not be able to do even `curl www.google.com` for example.

This will keep happening, and not only will SSH And GPG keys be the target, but any interesting data will be stolen.

And the problem is much larger than these typosquatting attacks. Abandoned Github projects taken over my malicious users, rogue Maven/npm/PyPI/what have you repositories, hacked accounts on any website that is used for distributing programs, feature branches in open source projecs that are automatically built on CI servers in side corporate networks, the possibilities to grab data and send it to somewhere on the internet are endless.

One security measure that somehow grew out of fashion over the last years, is at least on application servers, to disallow any outgoing network traffic, especially to the internet (at least any cloud environment I see nowadays allows it by default). This would largely prevent these sorts of attacks from being able to actually send anything out, but also prevent XXE attacks from happening, prevent reverse connections to an attacker host from being set up, make SSRF attacks harder to verify, and so on.

I strongly recommend whitelisting only the network traffic that your application actually needs.

Lots of problems are merely I/O related, which can be solved just fine with Python threads. As for number crunching (let's assume that means CPU intensive tasks), you can always still resort to multiprocessing (which can also be combined with multithreading, I fail to see why such paradigms cannot be combined).

I think it's important to be careful with drawing conclusions about what might or might not be used by whatever is running in your container and exposed to end users, if only for the simple fact that it might not be too obvious how a particular vulnerability gets exploited in the wild. CTF writeups often illustrate very nicely how there are quite creative ways to exploit vulnerabilities. It's not safe to assume you can predict every scenario. The path to an attack can be very complicated (but easy to execute, once the exploit is scripted).

I agree that the article doesn't emphasize what is the actual important point here (it is clickbait-ish), but the numbers they're presenting should (hopefully) trigger people to actually think about putting "continuous eyes" on the container images they're using. Just like you should continuously monitor your code, your application's dependencies, and your host system libraries.

A hacker needs to be right once, you need to be right 100% of the time. That's not marketing.

On premise deployed stacks are usually shielded off by network firewalls though, which cannot be said about most cloud based services (yes, I get it from a convenience point of view).

In order to attack an on premise application that is safely behind a firewall in a private LAN, one would have to come up with more creative, staged attacks (such as DNS rebinding or blind XSS attacks).

So, put simply, whereas the attack vector of a cloud based service is basically "the internet", this isn't usually the case with on premise deployments, and therefore they are inherently safer, or at least much harder to attack.

On top of that, cloud based services are usually multi tenant environments, so the assets to gain there are huge. One serious hole in such a service (be it Slack, logz.io, Github private repos, New Relic, etc.) will probably be a disaster.

But now you're hitting on one of the main points that was always criticized in PHP: the fact that it came with a lot of batteries included. The upside of this is that it makes PHP applications very easy to deploy, because often, you either don't need to resort to binary dependencies outside the realms of your OS's packaging system, or there will be a `php[major-version]-[library_binding]` package for your OS probably already in your OS's official package mirrors. Say all you want about PHP, but of all the application stacks I've had to deploy, PHP was and is probably still the simplest. The advantage of that shouldn't be underestimated. The only serious competitor in that space that I've ever touched is probably Go.

The article mentioned ctags, which actually gets you quite far in this regard. Combined with CtrlP (a vim plugin) and the Silver Searcher, I navigate large codebases quite quickly. I do agree that full blown IDEs probably have better support for refactoring code, but so far the advantages of using vim/tmux/unix commands are much more valuable to me:

- it's more natural for me to have a shell as a starting point and jump into an editor, rather than the other way around; it suits my workflow much better

- I can combine tools in a very convenient way, `:!ls /tmp` real quick from my vim, fuzzy search some json file hidden inside an atrocious directory structure, `:!git status -s` to see what I've changed, and so on

- IDEs are often language specific, which implies learning a new IDE for every language; I only need to know vim, which I use for every language I use but Java

- I'm more of a keyboard person rather than a mouse person, so terminal based tools that have intuitive keyboard bindings suit me better than IDEs which often require you to learn some really awkward key combinations to do things that I do very quickly in vim; just take a look at the cheatsheets for e.g. PyCharm/IntelliJ and you'll see what I mean (yes I know about the vim plugin for IntelliJ, but as an experienced vim user, it's not that great)