giulio

Sample porting of `core` to `Uint8Array` only here: https://github.com/lsd-cat/sigstore-js/commit/934a73a769b6ff68f1389b5d23b125cd033f657f

I have worked both on points 2 and 3, also porting the already existing tests to [vitest](https://vitest.dev/) as possible. At the moment, the Sigstore core module successfully builds with `vite`....

I am having a hard time verifying signatures :/ somehow the `verify()` from the Web Crypto API seems to always fail, there might be an issue in the input format,...

Fixed that, now all tests are passing using `vitest` and using browser-only native methods :D Does not mean it's fully compatible, but it's a good start!

Next step would be to port the `verify` package, and my understanding is that it depends on: * `core` -> we have it! * `bundle` -> if protobuf bindings works,...

> Yes, I think so. Specifically we want to have the client/app just send off normal HTTP requests instead of our current homegrown JSON protocol. But this is less important...

Unless we can come up with concrete, credible cases where AppArmor would save the day, I'd lean towards avoiding stacking up security protections if not proven necessary. > +1 for...

I was re-reading recent kernel vulnerabilities and techniques, namely - https://yanglingxi1993.github.io/dirty_pagetable/dirty_pagetable.html - https://pwning.tech/nftables/ Both very interesting reads, both targeting mostly temporal (rather than spatial) vulnerabilities (double free/UAF), both targeted at...

I think that something that uses the browser Crypto API and has been widely deployed in a browser context it what we should aiming for, both for maintainability, but also...

I also disagree with the idea that writing bindings and calling external libs is better for security. On the opposite, everything for which we write our custom bindings, is an...