zkSecurity
@zksecurityXYZ
Security audits, development, and research for ZKP, MPC, FHE, PQC, and more generally advanced cryptography. Contact us: hello@zksecurity.xyz
18 Following 7K Followers
We're proud to release MinaGuard, a native multi-sig wallet for Mina, together with @MinaFoundation 🪶
we're really lucky to be working with the best teams and most interesting projects in the space right now
It is difficult to explain just how impressive this is. The Tachyon Project (@TachyonZcash) and @zksecurityXYZ have raised the bar for high-assurance software.
zkSecurity audits formal verifications, the line is this way ⬇️
We hired @zksecurityXYZ to audit Ironwood's formal verification. Rather than looking for security bugs, they tried to *add* bugs to see if the formal verification could catch them!
Here's our summary, along with the report:
Show more
zkao v3 is coming out soon (our AI bugfinder for cryptographic code)! If you haven't been approved on the platform ( shoot me a message. We are opening up applications to more projects as we continue our slow roll out.
Show more
⚠️ Important one for anyone building on @MinaProtocol
o1js 3.1.0 is out, with fixes for three security vulnerabilities identified with the help of @zksecurityXYZ .
If you're building a zkApp, upgrading should probably be top of the list:
npm install o1js@3.1.0
1/3
Show more
Nine months ago, at a workshop about formal mathematics and AI, I heard about formalizing the whole mathematics in six months. I took it as half-joke. I wanted to record this. Might be useful for estimating the current joke-to-reality velocity.
Show more
As everyone is overwhelmed with bug reports, why not use a tool that's low on false positives? Use
zkML still feels impractical for broad use cases
But in Session 11 of Proof is in the Pudding, @cryptodavidw explains why it’s less ridiculous than it seems, and walks us through what’s needed for it to become a more widely useful technology
⌛️ TIMESTAMPS
0:00 Intro
1:44 What is zkML
5:32 Inference vs training
7:52 Alternatives to ZK for verifiable inference
10:44 Inside a transformer
15:25 Proving transformer computation
27:12 Quantization & SNARK-friendly models
31:58 KV caching & optimization opportunities
35:08 Agent-to-agent proofs & dynamic circuits
37:18 How zkML scales
38:50 Market for verifiable ML
Proof is in the Pudding is produced in collaboration with @zksecurityXYZ
Show more
We released a new whiteboard session on zkML: proving inference of LLMs using zero-knowledge proofs, in collaboration with @archetypevc
why use zkao and not do it yourself?
because we have a team dedicated on working on this tool full time
because we continuously improve zkao, teaching it to find bugs that it couldn't find in audits
because we aggregate all the models and techniques so you don't have to
Show more
My life is dope and I do dope shit
Words cannot describe how satisfying it was to finally close this from propext, Classical.choice, Quot.sound.
Did you know that Clean has a website and a pragmatic introduction to formal verification for zk devs?
We’re happy to support the proximity prize challenge efforts, now with machine-verifiability!
currently hosts two challenges, to improve what we know about the security bounds of proximity gaps. Since the challenges are expressed in Lean, every accepted result carries a strong guarantee, in the same style as
This follows the series of efforts in the industry and in our team to utilize verifiability for fearless optimization.
Starting from a security bits baseline of 53 (lower bound) and 128 (upper bound), AIs (and people too) will collaborate to improve the state of the art, and hopefully interesting ideas that can advance science will rise out of it!
Show more
A new golf course has opened on zkGolf! ⛳️
The new challenges we just added are hashes over *binary fields*: SHA2 and KangarooTwelve
Aaaand we have reached SOTA on (emulated) fixed-based scalar multiplication as well:
Thanks to and all the other golfers that made this possible.
