Zero Cool

65 posts

Zero Cool banner
Zero Cool

Zero Cool

@ZeroCool_AI

AI + human hybrid systems for the cyborg era.

Katılım Temmuz 2025
2 Takip Edilen955 Takipçiler
Zero Cool
Zero Cool@ZeroCool_AI·
This looks basic, but we still see protocols falling for this in 2026. Can you tell why Zero Cool panicked over these lines of Code?
English
0
0
8
448
Zero Cool
Zero Cool@ZeroCool_AI·
In the recent VII-Finance-Contracts contest on Cantina, Zero Cool found a very interesting issue in which deterministic wrapper salts used across different create methods allow early pools to permanently block alternative topologies. This happened because multiple create functions reuse the same salt derivation, even though those salts represent different logical flows. Full write-up to be posted soon!
English
1
1
15
1.3K
Zero Cool
Zero Cool@ZeroCool_AI·
❤️❤️❤️
Zero Cool tweet media
QME
1
0
11
281
Zero Cool
Zero Cool@ZeroCool_AI·
That's tough, but Zero Cool always grabs it by the neck.
Zero Cool tweet media
English
0
0
3
192
Zero Cool retweetledi
Artem Chystiakov
Artem Chystiakov@Arvolear·
Wanna give a shoutout to @ZeroCool_AI because I'm genuinely shocked by the depth of the findings. Guys just randomly DMed me (I didn't pay for anything) a report with insane beta. Read the full audit here: hackmd.io/rAhJFzFmTgeDVK…
English
2
3
33
3.5K
Zero Cool
Zero Cool@ZeroCool_AI·
Late last year, we ran Zero Cool on the Solarity lib and identified 2C, 4H, 6L. We will share some of the very interesting findings soon.
English
0
0
5
297
Zero Cool
Zero Cool@ZeroCool_AI·
Not all initialization counts, especially under proxy patterns. In this finding, we establish that inline initialization is just as ineffective as constructor initialization in a UUPS implementation contract: neither ever reaches the proxy's storage. Zero Cool spotted that gap on a critical variable during this review.
Zero Cool@ZeroCool_AI

x.com/i/article/2049…

English
0
0
6
736
Zero Cool
Zero Cool@ZeroCool_AI·
Zero Cool definitely ate at the Mento-V3 contest, as we smashed through to 2nd place. Full write-ups of our findings are coming through.
Zero Cool tweet media
English
0
1
16
724
Zero Cool
Zero Cool@ZeroCool_AI·
Multiple new BB findings confirmed - more info coming soon!
Zero Cool tweet media
English
0
1
9
283
Zero Cool
Zero Cool@ZeroCool_AI·
One obvious takeaway from this finding: Zero Cool has a remarkable understanding of how ZK circuits should work and an eye for spotting the gaps that truly matter. This time, the missing piece was a single perform_memcopy flag inside a selector-gated range check, enough to let non-canonical 16-bit limbs slip through in ZKsync’s bigint handling. More to come!
Zero Cool@ZeroCool_AI

x.com/i/article/2048…

English
0
0
14
921