고정된 트윗
Kai
5.2K posts
Darkdrop invented the credit note model a new privacy architecture where the claim tx shows zero amounts, zero SOL movement, zero inner instructions. no other mixer does this.
Then we added the note pool, a second layer that mixes all credit notes together regardless of amount. two ZK proofs to break. anonymity set is everyone.
open source. live on devnet. built solo.
darkdrop.app
English
Hey @colosseum Frontier builders!
I’m a judge. I’m even billed above @toly. Which means I’m probably more important.
Why are you not in my replies shilling your project, asking for feedback?!?
English
Been on the business plan since last year 😅 and not a single like or RT from you, ser.
Proud Helius business user btw.
mert@mert
i've angel invested in more Solana projects than anyone alive (including toly) from now on, if you get a @Helius business or professional plan, I will give you a complimentary pitch deck review as well as intros nothing more expensive than not using Helius on Solana
English
First fork just landed on the #darkdrop repo. @colosseum hackathon teammate onboarded and already studying the frontend flow.
Open source privacy protocol with a real contributor now reviewing the codebase. this is how it starts.
Two layers of ZK proofs. nine instructions. three circuits. and now more than one set of eyes on the code.
github.com/hitman-kai/dar…
English
The primitives aren't as limited as people think. there's room for genuinely new stuff if you actually study what exists and find the gaps instead of forking what's already there.
if anyone wants to look under the hood: github.com/hitman-kai/dar…
brimigs@b_migliaccio
I agree, but even getting to something like an onchain Nasdaq happens faster with more builders exploring the space even if primitives are limited today, more motivated builders = higher chance of new ones emerging & even with a limited variation, theres enough interesting tech to nerdsnipe a lot of bright minds.. if communicated well enough IMO more perspectives collaborating = more innovation
English
This is exactly what pushed me to build the credit note model. looked at every privacy primitive on solana and ethereum they all leak amounts in the claim tx. same architecture recycled with different frontends.
so i built a new one. two-phase claim where the proof tx shows zero amounts and the withdrawal uses direct lamport manipulation instead of CPI transfers. then added a second mixing layer on top of it because one layer still has amount correlation.
the primitives aren't as limited as people think. there's room for genuinely new stuff if you actually study what exists and find the gaps instead of forking what's already there.
open source if anyone wants to look under the hood: github.com/hitman-kai/dar…
English
I agree, but even getting to something like an onchain Nasdaq happens faster with more builders exploring the space
even if primitives are limited today, more motivated builders = higher chance of new ones emerging
& even with a limited variation, theres enough interesting tech to nerdsnipe a lot of bright minds.. if communicated well enough
IMO more perspectives collaborating = more innovation
English
Ok hear me out:
There are over 2 million monthly active Rust devs and about 4k monthly active solana devs
If we could just tap into 1% of the current rust developer community, the growth of Solana would probably be exponential.
Lets brainstorm, how can we achieve this?
English
Kai 리트윗함
right now withdrawals are full amount you withdraw everything in the credit note at once.
Partial withdrawals with hidden change are next.
the architecture supports it a new circuit proves withdraw_amount + change_amount = original_amount with range checks, and the change becomes a fresh unlinkable credit note.
Same Poseidon commitment scheme, same re-randomization. but honestly the note pool already weakens amount matching significantly because all amounts sit in the same pool regardless of size.
The observer can't filter by amount anymore, the anonymity set is everyone, not just same-amount deposits. partial withdrawals will break the 1:1 correlation completely. coming soon.
fam.work 🩵@famdotwork
@HitmanNoLimit the amount-matching attack is the obvious one and it's wild how many mixers just... ignore it. credit note model is clean for this. how are you handling note fragmentation when someone wants to withdraw a non-standard amount?
English
1 layer of privacy isn't enough. every mixer on solana shows the amount in the claim tx. match deposit to withdrawal and privacy is gone.
so we built two layers. the credit note model hides the claim. the note pool mixes everything together regardless of amount. two ZK proofs. zero linkable data.
no one has deployed this before.
darkdrop.app
6wKRRP1c2gkWESch723bmgCWiBYiYvn4krspZXdApump
English
Every privacy protocol on Solana has one layer of ZK proofs.
@darkdrop_sol now has two.
layer 1 hides which deposit you claimed. layer 2 hides which credit note you're redeeming. an observer needs to break both to link sender to receiver.
and the second layer mixes all amounts together 0.1 SOL and 10 SOL in the same pool, indistinguishable. your anonymity set is no longer limited to same-amount deposits. it's everyone.
no chain has this deployed. we just shipped it on Solana.
github.com/hitman-kai/dar…
English
Every HIGH and MEDIUM finding is fixed. The credit note model of $darkdrop now has:
Zero amounts in claim TX
Zero SOL movement in claim TX
Zero inner instructions on withdrawal
Zero event linkage (M-03-NEW fixed)
Zero account data linkage (M-01-NEW fixed with salt)
Rug-pull protected (H-01-NEW fixed with obligation tracking)
Fee diversion blocked (H-01, M-01 fixed)
The protocol is bulletproof on the security side.
verify here
github.com/hitman-kai/dar…
English
Security audits for $darkdrop
DarkDrop@darkdrop_sol
3 security audits completed. all public. all findings documented. all critical fixes deployed. 👇
English
Bring back holding culture. $darkdrop
DarkDrop@darkdrop_sol
Gm! Let's bring back the holding culture. $darkdrop
English
Kai 리트윗함
Introducing Project Glasswing: an urgent initiative to help secure the world’s most critical software.
It’s powered by our newest frontier model, Claude Mythos Preview, which can find software vulnerabilities better than all but the most skilled humans.
anthropic.com/glasswing
English