Zellic ✓
@zellic_io · joined 16 Nov 2021
Security reviews and research that keep winners winning. We apply unmatched hacking talent to secure critical software for the most innovative teams.
18 144Followers
18Following
1 735Posts total
1.9MViews on collected posts
Derniers posts
Recherche de nouveaux posts…
One permissionless call and your deposit is… miner fees?
Legacy P2PKH signatures, trusting user calldata, and a signer with no context of what it’s signing.
Here’s how 64 missing bits allowed @RelayProtocol’s permissionless `sweep()` to drain Bitcoin funds to fees. 🧵 https://t.
3.9K views · 34 likes
· 6 reposts · 1 replies
Open on X →
We're very excited to be sponsoring Paged Out! once again! Check out Issue #9👇 https://t.co/DxicgzO096
6.5K views · 26 likes
· 3 reposts · 1 replies
Open on X →
Paged Out! #9 is live!
Read, share, enjoy!
https://t.co/kVmdLe0l7s
We are running final QA on all PO! issues that should be available on Lulu (https://t.co/RoJJnVAYnY) in a few days!
In case you want to support us, we have a Patreon now - https://t.co/CkLKpl1WWU https://t.co/
14.3K views · 74 likes
· 28 reposts · 2 replies
Open on X →
Note that all of our proofs are valid proofs for the statements they claim to prove. The issue lies in those statements not meaning what they were intended to mean.
While our circuits were still sound and complete, the flaw in the spec for mainCost allowed us to break the larger
30.9K views · 70 likes
· 7 reposts · 0 replies
Open on X →
The cost claim only covers the first branch and is thus trivial to prove. https://t.co/P91iba8HHb
2.1K views · 10 likes
· 0 reposts · 1 replies
Open on X →
Both branches are sound and complete for their respective inputs, so we can also prove soundness and completeness for our `main`: https://t.co/4nG0k5H9Rj
1.6K views · 10 likes
· 0 reposts · 1 replies
Open on X →
Our solution thus branches on whether the input is equal to this default constant.
If it is, we just return the correct result as a constant, requiring zero allocations or constraints.
If it is not, we fall back to the baseline circuit. https://t.co/SE4uLhsTi9
1.6K views · 11 likes
· 0 reposts · 1 replies
Open on X →
So what is the default input?
Ultimately, the input boils down to consisting of `Expression`s.
These can be a constant, a symbolic variable, or a polynomial in other expressions. https://t.co/5RlKkYO4Q2
1.8K views · 11 likes
· 0 reposts · 1 replies
Open on X →
The default expression turns out to be the constant zero.
The cost counting function thus gets passed `main` applied to an input consisting only of constants with value zero. https://t.co/NzTVc3WfAg
1.6K views · 10 likes
· 0 reposts · 1 replies
Open on X →
In this screenshot you can see the types of everything we need to provide.
The circuit is provided with `main`. This takes the circuit input as an argument, and returns a circuit (constraint system).
Note that `main` having this type means that the constraints are allowed to ht
2.3K views · 14 likes
· 0 reposts · 1 replies
Open on X →
Solutions claim a number of allocations and constraints, and must provide a proof of `mainCost`, which is intended to be the statement that `the` circuit allocates that many witnesses and has that many constraints.
But note that the argument to `Challenge.CostR1CS.circuitCount`
2K views · 14 likes
· 0 reposts · 1 replies
Open on X →
How does it work?
You upload a circuit, along with a Lean 4 proof of correctness, and submissions are ranked based on circuit cost, which is the sum of the number of witness allocations and the number of constraints.
As all submissions are formally verified, developers could ht
2.9K views · 18 likes
· 0 reposts · 1 replies
Open on X →
You "win" https://t.co/y2fdv7oJiN – ZKSecurity’s FV contest – by writing the most optimized formally verified circuit for cryptographic primitives.
But how does the best SHA256 circuit have a cost of 0?
Here's how we exploited the FV spec for the top score on all circuits.👇 htt
49K views · 147 likes
· 10 reposts · 6 replies
Open on X →
We formally verified a Plonky2 gate.
ZK proofs show circuit satisfaction, but do they prove the statement you rely on?
One flaw and attackers can convince a verifier that false statements are true.
We closed that gap with formal verification.
Here's how we proved it in Lean.
7.7K views · 97 likes
· 10 reposts · 5 replies
Open on X →
How to find a $65,000 zero-day in Chrome V8:
Meet @eternalsakura13, researcher at Zellic.
- Top 3 Chrome VRP 2022–2024
- Top 2 Facebook whitehat in 2023
- Top 10 MSRC MVR in 2025
Here’s a walk through the mind of one of the world’s best Chrome researchers. Can you follow along?
55.2K views · 510 likes
· 57 reposts · 4 replies
Open on X →
Solana’s VM has bugs.
Or to be more precise, it had these two bugs:
- One bug leaked memory.
- Another corrupted “untouchable” memory.
Both were found in Solana’s high-performance sBPF engine.
These bugs have since been fixed.
But the following bugs reveal what happens when
10.6K views · 101 likes
· 7 reposts · 5 replies
Open on X →
You’re probably using WebViews wrong.
There are a million ways to use a WebView wrong. Properly securing a WebView is hard.
In this thread, we’ll cover common vulnerabilities in wallet WebView implementations and the ways to properly secure WebViews. https://t.co/DG6a4fK7VG
25.7K views · 241 likes
· 37 reposts · 1 replies
Open on X →
ZK protocols use special hash functions that are algebraically simple.
But if our proof system supported lookup tables, could we do better?
Introducing Polocolo: a new ZK-friendly hash function for PlonKup, co-authored by Zellic Cryptographer @baaaaaarkingdog. https://t.co/hviZ
8K views · 73 likes
· 9 reposts · 5 replies
Open on X →
BLS signatures are everywhere, from Ethereum’s consensus to EigenLayer. But it’s easy to use them wrong.
What are BLS signatures? Let’s talk about the right way and the wrong way to use them: https://t.co/5oQ44rhSyP
54.4K views · 330 likes
· 49 reposts · 5 replies
Open on X →
Last month, @chrisling_dev introduced WHLUSDC, a Hyperliquid-native stablecoin.
On a Spaces after the announcement, Chris mentioned a bug purposely introduced in WHLUSDC as a challenge.
Our team quickly got to work.
This is how we found the bug along with the POC to prove it.
18.3K views · 84 likes
· 10 reposts · 2 replies
Open on X →
Face aux comptes de taille comparable
18 posts des 90 derniers jours, à côté de la tranche de 10K–100K abonnés. diffusé largement, mais peu de ces spectateurs réagissent.
Vues médianes2 012ce compte949médiane pour 10K–100K
Portée, %11.09%ce compte3.31%médiane pour 10K–100K
Engagement, %0.66%ce compte1.88%médiane pour 10K–100K
| Indicateur | Ce compte | Médiane pour 10K–100K | Rapport |
|---|---|---|---|
| Vues médianes par post | 2 012 | 949 | 2.12× |
| Portée (vues ÷ abonnés) | 11.09% | 3.31% | 3.35× |
| Taux d'engagement | 0.66% | 1.88% | 0.35× |
Autres comptes de cette tranche → Comparer avec un autre compte → Comment ces repères sont établis →
Growth & engagement
How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.
Views per post
2K3 Jul
2.3K
1.6K
1.8K
1.6K
1.6K
2.1K
30.9K
860
7376 Jul
14.3K27 Jul
6.5K
3.9K27 Aug
20028 Aug
Last 14 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.
Engagement rate per post
0.76%3 Jul
0.65%
0.67%
0.68%
0.75%
0.68%
0.53%
0.26%
2.33%
0.27%6 Jul
0.79%27 Jul
0.46%
1.07%27 Aug
0.00%28 Aug
Reactions — likes, reposts, replies and quotes — divided by views. Median for 10K–100K accounts is 1.88%.
What the audience does
Likes79.4%1 911 in total
Reposts9.7%233 in total
Replies2.3%55 in total
Quotes1.7%42 in total
Bookmarks6.9%166 in total
Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.
The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.
Comptes similaires
DIRECTOR PINK ✓
@directorpink_
36.3K followers
· 1.9K posts
Shridhar V
@iimcomic
36.3K followers
· 10.8K posts
Frank Mugisha ✓
@frankmugisha
36.3K followers
· 41.6K posts
Dave Parker ✓
@iamdaveparker
36.3K followers
· 180.6K posts
Minsleif ✧ Leify Lore Keeper ✓
@Minsleif
36.3K followers
· 13.4K posts
سهيل أبانمي ✓
@abanmis
36.3K followers
· 5.4K posts
Chris Hummer ✓
@chris_hummer
36.3K followers
· 36K posts
IG ✓
@IGcom
36.3K followers
· 34.1K posts