I think it’s more like: a script running in your browser tab, given some untrusted input (say, an image you import into an editing program) could be compromised by malicious input, such that the script runs malicious code and exfiltrates other data from the app (say, your email address or other private data you used on the page).

That's a definite risk in general.

But worth adding that Wasm by design lets you statically inspect all the things the Wasm can call. If the Wasm receives access to networking, or eval(), etc., then there is a real risk. However, often the Wasm imports and exports are extremely limited and easy to verify for safety.

Section 5.1 gives a good example: an image sharing service that processes images using libpng. This has to be combined with a document.write() of HTML in C, which is realistic.

Someone who controls the image data can perform an XSS, i.e. steal your credentials for the website, or your credit card info if that's stored server-side. That's not as valuable a target as controlling your computer, but it's not nothing, and can be chained.

This is completely realistic as Figma is full blown image editor written in C++ compiled to WebAssembly :)

https://www.figma.com/blog/webassembly-cut-figmas-load-time-...

"Someone who controls the image data" to me that sound like something outside the sandbox manipulating the image data given to the sandbox.

If this is the case it mean there is already a hacker or virus running outside the sandbox which mean I already have bigger problem to worry about :).

No, it means a random user can upload a malicious image and hack the service.

neither, the program might just crash or be buggy. Maybe in some rare cases do malicious things to the programs running in the sandbox.