quickjs-emscripten VS SES-shim

Based on your comment below I think you figured out the difference - but if you're looking to execute JS, you can pick between ShadowRealm (where available, or using a polyfill) or my library quickjs-emscripten.

Pros of quickjs-emscripten over ShadowRealm:

- You can use quickjs today in any browser with WASM. ShadowRealm isn't available yet, and polyfills have had security issues in the past. See https://www.figma.com/blog/an-update-on-plugin-security/

- In ShadowRealm eval, untrusted code can consume arbitrary CPU cycles. With QuickJS, you can control the CPU time used during an `eval` using an [interrupt handler] that's called periodically during the eval.

- In ShadowRealm eval, untrusted code can allocate arbitrary amounts of memory. With QuickJS, you can control both the [stack size] and the [heap size] available inside the runtime.

- quickjs-emscripten can do interesting things with custom module loaders and facades that allow synchronous code inside the runtime to call async code on the host.

Pros of ShadowRealm over QuickJS:

- ShadowRealm will (presumably?) execute code using your native runtime, probably v8, JavaScriptCore, or SpiderMonkey. Quickjs is orders of magnitude slower than JIT'd javascript performance of v8 etc. It's also slower than v8/JSC's interpreters, although not by a huge amount. See [benchmarks] from 2019.

- You can easily call and pass values to ShadowRealm imported functions. Talking to quickjs-emscripten guest code requires a lot of fiddly and manual object building.

- Overall the quickjs(-emscripten) API is verbose, and requires manual memory management of references to values inside the quickjs runtime.

[interrupt handler]: https://github.com/justjake/quickjs-emscripten/blob/main/doc...

[stack size]: https://github.com/justjake/quickjs-emscripten/blob/main/doc...

[heap size]: https://github.com/justjake/quickjs-emscripten/blob/main/doc...

[benchmarks]: https://bellard.org/quickjs/bench.html

The thing I want to achieve with WebAssembly is still proving a lot harder than I had anticipated.

I want to be able to take strings of untrusted code provided by users and execute them in a safe sandbox.

I have all sorts of things I want this for - think custom templates for a web application, custom workflow automation scripts (Zapier-style), running transformations against JSON data.

When you're dealing with untrusted code you need a really robust sandbox. WebAssembly really should be that sandbox.

I'd like to support Python, JavaScript and maybe other languages too. I want to take a user-provided string of code in one of those languages and execute that in a sandbox with a strict limit on both memory usage and time taken (so I can't be crashed by a "while True" loop). If memory or time limit are exceeded, I want to get an exception which I can catch and return an error message to the user.

I've been exploring options for this for quite a while now. The furthest I've got was running Pyodide inside of Deno: https://til.simonwillison.net/deno/pyodide-sandbox

Surprisingly I've not found a good pattern for running a JavaScript interpreter in a WASM sandbox yet. https://github.com/justjake/quickjs-emscripten looks promising but I've not found the right recipe to call it from server-side Python or Deno yet.

Can Extism help with this? I'm confident I'm not the only person who's looking for a solution here!

"Safely execute untrusted Javascript in your Javascript, and execute synchronous code that uses async functions" quickjs-emscripten, NPM

This maybe, as a start?

https://github.com/justjake/quickjs-emscripten

QuickJS Running in WebAssembly\ (17 comments)

The library was inspired by Figma’s blog posts about their plug-in system: https://github.com/justjake/quickjs-emscripten#background

If you need to call into user-generated Javascript synchronously or have greater control over the sandbox environment, you can use WebAssembly to run a Javascript interpreter: https://github.com/justjake/quickjs-emscripten#quickjs-emscr...

QuickJS in WebAssembly is much slower than your browser's native Javascript runtime, but possibly faster than async calls using postMessage. As an added bonus, it can make async functions in the host appear to be synchronous inside the sandbox using asyncify: https://emscripten.org/docs/porting/asyncify.html.

You can use One Weird Trick with generator functions to make your code "generic" over synchronicity. I use this technique to avoid needing to implement both sync and async versions of some functions in my quickjs-emscripten library.

The great part about this technique as a library author is that unlike choosing to use a Promise return type, this technique is invisible in my public API. I can write a function like `export function coolAlgorithm(getData: (request: I) => O | Promise): R | Promise`, and we get automatic performance improvement if the user's function happens to return synchronously, without mystery generator stuff showing up in the function signature.

Helper to make a function that can be either sync or async: https://github.com/justjake/quickjs-emscripten/blob/ff211447...

Uses: https://cs.github.com/justjake/quickjs-emscripten?q=yield*+l...

This seems like a pretty nice, recently enabled way of getting a sandboxed js environment: QuickJS compiled to WASM: https://github.com/justjake/quickjs-emscripten.

I used E in the 90s: http://erights.org/

I haven't kept up with newer systems but I've heard of https://github.com/endojs/endo and just came across http://reports-archive.adm.cs.cmu.edu/anon/home/anon/isr2017... (which says "in the style of the E programming language" -- that's as far as I've read) while looking that up.

WebAssembly was designed to follow the same capability security principles. CHERI too as someone else just brought up.

There are other potential solutions I haven’t explored close enough (like Endo and SES), or completely omitted as they’re based on an imperfect blacklist-based approach to security (like sandboxed WebWorkers). However, the mentioned 4 solutions are the top contenders, at least in my mind.

I don't know why you are being silently downvoted, as I think it is worth talking about the potential of using static analysis to improve things.

One promising approach is Endo[0] which "uses LavaMoat to automatically generate reviewable policies that determine what capabilities will be distributed to third party dependencies."

[0] https://github.com/endojs/endo

Agoric moved forward and Realms gave way to SES

https://github.com/endojs/endo/tree/master/packages/ses

And Endo is a set of tools (being) built around it to make it more practical for particular usecases

Yea you could restrict the app by whitelisting only the network services and folders that it will use and that's pretty valuable though at least on Linux could already easily be achieved otherwise. It's good that Deno makes it easy but let's be honest, most people will just pass -A.

I'd love to see a permissions system on a library basis. It would ask the first time a dependency is added and when a new permission is requested after an update. Javascript doesn't make that easy though by being so dynamic. SES could maybe help: https://github.com/endojs/endo/blob/master/packages/ses/READ...

I was poking around on the internet a bit earlier and I found this project. It looks pretty cool, and I figured perhaps a few of y'all might find it cool too!

I have no idea if it actually sandboxes networking by default. This other project, endo[0], seems to add some of that functionality.

Regardless of the maturity though, it makes me excited to see this type of work getting done now!

(What made me want to research it was this[1] thread from the other day.)

0: https://github.com/endojs/endo

1: https://news.ycombinator.com/item?id=30215212

Fortunately the problem could become more tractable if something like SES / Endo takes off:

"Endo protects program integrity both in-process and in distributed systems. SES protects local integrity, defending an application against supply chain attacks: hacks that enter through upgrades to third-party dependencies. Endo does this by encouraging the Principle of Least Authority. ... Endo uses LavaMoat to automatically generate reviewable policies that determine what capabilities will be distributed to third party dependencies."

https://github.com/endojs/endo

Yeah. JavaScript is probably the closest to being there (with things like SES[0], LavaMoat[1], etc.) but we're not quite there yet. It's just shocking that this sort of thing is as seemingly obscure as it is; it's like the whole industry has collectively thrown up their hands and said code execution is unavoidably radioactively dangerous. (While simultaneously using package managers that... well.) But it doesn't have to be!

[0] https://github.com/Agoric/ses-shim

[1] https://github.com/LavaMoat/LavaMoat