slsa VS language

Short answer: not where it counts.

My work focuses on recognizing known functions in obfuscated binaries, but there are some papers you might want to check out related to deobfuscation, if not necessarily using ML for deobfuscation or decompilation.

My take is that ML can soundly defeat the "easy" and more static obfuscation types (encodings, control flow flattening, splitting functions). It's low hanging fruit, and it's what I worked on most, but adoption is slow. On the other hand, "hard" obfuscations like virtualized functions or programs which embed JIT compilers to obfuscate at runtime... as far as I know, those are still unsolved problems.

This is a good overview of the subject, but pretty old and doesn't cover "hard" obfuscations: https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=1566145.

https://www.jinyier.me/papers/DATE19_Obf.pdf uses deobfuscation for RTL logic (FGPA/ASIC domain) with SAT solvers. Might be useful for a point of view from a fairly different domain.

https://advising.cs.arizona.edu/~debray/Publications/generic... uses "semantics-preserving transformations" to shed obfuscation. I think this approach is the way to go, especially when combined with dynamic/symbolic analysis to mitigate virt/jit types of transformations.

I'll mention this one as a cautionary tale: https://dl.acm.org/doi/pdf/10.1145/2886012 has some good general info but glosses over the machine learning approach. It considers Hex-rays' FLIRT to be "machine learning", but FLIRT just hashes signatures, can be spoofed (i.e. https://siliconpr0n.org/uv/issues_with_flirt_aware_malware.p...), and is useless against obfuscation.

Eventually I think SBOM tools like Black Duck[1] and SLSA[2] will incorporate ML to improve the accuracy of even figuring out what dependencies a piece of software actually has.

[1]: https://www.synopsys.com/software-integrity/software-composi...

[2]: https://slsa.dev/

The dependency chain is certified! SLSA!

The things you mentioned are not solved by a typical "SBOM" but e.g. CycloneDX has extra fields to record provenance and pedigree and things like in-toto (https://in-toto.io/) or SLSA (https://slsa.dev/) also aim to work in this field.

I've spent the last six months in this field and people will tell you that this or that is an industry best practice or "a standard" but in my experience none of that is true. Everyone is still trying to figure out how best to protect the software supply chain security and things are still very much in flux.

It's multi-pronged and I imagine adopters may use a subset of features. Broadly, I think folks are going to be interested in a) branch/tag/reference protection rules, b) file protection rules (monorepo or otherwise, though monorepos do pose a very apt usecase for gittuf), and c) general key management for those who primarily care about Git signing.

For those who care about a and b, I think the work we want to do to support [in-toto attestations](https://github.com/in-toto/attestation) for [SLSA's upcoming source track](https://github.com/slsa-framework/slsa/issues/956) could be very interesting as well.

It doesn't have to be. Corporations which are FedRAMP[1] compliant, have to build software reproducibly in a fully isolated environment, only from reviewed code.[2]

[1] https://en.wikipedia.org/wiki/FedRAMP

[2] https://slsa.dev/

SLSA stands for Supply chain Levels for Software Artifacts, and it is a framework that aims to provide a set of best practices for the software supply chain, with a focus on OSS. It was created by Google, and it is now part of the OpenSSF. It consists of four levels of assurance, from Level 1 to Level 4, that correspond to different degrees of protection against supply chain attacks. Our CTO Paolo Mainardi mentioned SLSA in a very good article on software supply chain security, and we also mentioned it in another article about securing OCI Artifacts on Kubernetes.

SLSA.dev

I noticed that I can enable inline-class as an experiment to play with Extension Types. You need to also add sdk: ^3.3.0-0 to your pubspec.yaml.

Current syntax is not all that bad if you are going to do OO and add various helper methods on `Message` and its subclasses, but if you just want to define your data and no behavior / helpers - then it is exceedingly verbose.

[1]: https://github.com/dart-lang/language/issues/3021

Reference

A future version of Dart will probably support macros which should make this all a bit easier to use, similar to how Swift 5.9 works which makes already fantastic use of its new macro capabilities by integrating mobx (or solidjs) like reactivity into SwiftUI by a harmlessly looking @Obervable annotation.

Coming from a Dart context here where that team is also looking at adding Macros to the language. It was really interesting to compare and contrast some of the approaches https://github.com/dart-lang/language/blob/main/working/macr...

Yes, that needs a bit of boilerplate for the constructor declaration and the extra build method, but I personally don't mind and with implicit constructors this will become much easier. Also, you get a performant UI as Flutter knows to not redraw widgets that didn't change.

I know that it would be nice not to use the generator at all, but we have to wait until static metaprogramming is implemented in dart. https://github.com/dart-lang/language/issues/1482

I implemented a prototype version of the algorithm in that paper when exploring exhaustiveness checking for pattern matching in Dart.

I found it pretty easy to understand, but also really easy to get it to generate huge combinatorially large spaces. Some careful memoization and deduplication helped, but even so I never got the performance to a state I considered acceptable.

Instead, I went with Luc Maranget's classic approach and figured out a way to adapt it to a language with subtyping (with a ton of work from Johnni Winther to figure out all of the hard complex cases around generics):

https://github.com/dart-lang/language/blob/main/accepted/fut...

The performance (in the prototype!) was dramatically better. You can always make pattern matching go combinatorial, but I haven't seen any real-world switches get particularly slow with our approach yet, and we have some fairly large tests of matching on tuples of enums.