[Nzen]

I think that this is a plugin library for teams that want to offer a platform for the public (or an LLM-AI) to submit code to. If your team writes some code, you don't generally sandbox it from yourself, you just amend your program: you don't need a sandbox. But, if you want to run code that you don't trust, you should run it in a way that prevents it from causing problems if it is actually dangerous (like a virus or accidentally overwrites your files with blank files). That's what a sandbox like kyushu promises to do.

So, with a sandbox library like this, you could - say - write a website that hosts games (like itch.io or newgrounds) that hosts games on the world wide web. The sandbox part can give you confidence that, if a villain's programmer henchmen uploads a virus instead of a game, it can't infect your platform or other games on the website. Or, if a LLM-AI written game is accidentally tries to take up all the memory of the computer, it can't ask the operating system for more than is in the sandbox.

[zuzululu]

how is it different from firecracker or other containerization ? what makes it secure enough to make those claims?

[nevon]

Firecracker launches small, but otherwise general purpose virtual machines. Containers, at least the standard implementations that most of us use, use kernel features like namespaces to isolate workloads, but still share a kernel so the sandboxing is not as strong.

Wasm is a virtual machine, just like for example the jvm is, that is designed around only allowing the executed program access to the host runtime via specific apis that are subject to security policies. It does not run arbitrary software, but rather only software built to target specifically wasm.

The software this post is about is just bundling a wasm runtime with other software for convenience.