I Found 150+ Vulnerabilities in DeFi Protocols. Here's Why I Can't Do Anything About It.
A few months ago I started building a tool. Nothing fancy at first — just wanted to see how far you could push automated smart contract analysis. Static scanners, custom detectors, on-chain verification, AI-assisted deep dives. Iterated on it for weeks.
Then I pointed it at live DeFi protocols.
150+ verified vulnerabilities. 40+ protocols. Several criticals with working proof-of-concept exploits tested against mainnet forks. Not "theoretical under certain conditions." Working. Verified. Real money at risk right now.
And here's where the story gets dark.
— -
Try reporting a bug without revealing who you are.
I'm serious. Just try.
Immunefi — passport, selfie with ID, full identity verification. Cantina — same. Sherlock for large payouts — same. HackerOne — don't even get me started.
Okay, fine. Direct channels then. GitHub Security Advisories. Security emails. Discord. I tried. Some teams respond fast and professionally — genuinely impressed when that happens. Others go silent for weeks. Some auto-reply with "please submit through Immunefi." Some don't have a public security contact at all — good luck finding one.
Meanwhile the bugs sit in production code. Meanwhile real users have real money in these contracts.
— -
Here's what kills me about this.
These protocols raised $50–100M in funding rounds. They run bug bounty programs with caps at $1M, $3M, $15M. They tweet about their "commitment to security" after every major hack in the space.
But an independent researcher who finds a real vulnerability before a hacker does — has to either fully doxx themselves to a centralized intermediary, or walk away with nothing.
One protocol I found — won't name it — had a bug that would brick every single vault simultaneously. All of them. At once. Nobody deposits, nobody withdraws. I reported it.
Response: "We have a program on Cantina, please submit there."
Cantina requires KYC I can't pass.
The bug is still in the code.
— -
This isn't a bug in the system. It's a feature.
KYC on bug bounty platforms exists to protect platforms from regulators and protocols from uncomfortable questions. User safety is a marketing talking point, not an actual priority.
An industry built on trustless, permissionless infrastructure — requires you to hand over your passport to report that someone else's code is broken.
Let that sink in.
— -
There are people who get it.
Hats Finance. Small platform, not the most well-known — but the only major one where everything works the way Web3 is supposed to work. Fully on-chain. Encrypted submissions. Crypto payouts directly to wallet. No documents. No intermediaries. Just you, the protocol, and a smart contract holding the bounty.
Why isn't this the standard?
Because large protocols need compliance more than they need security. That's the honest answer.
— -
What I'm doing with these findings.
Continuing to push disclosures through every available direct channel. GitHub Security Advisories, direct emails, official security contacts. Some teams have already been notified. The rest will be.
90 days. That's the Google Project Zero standard. Report the bug, give the team time to fix it. If 90 days pass with no response or no fix — full public disclosure. Technical details included. This isn't a threat, it's just how responsible disclosure has worked for twenty years.
User funds matter more than a protocol's comfort with admitting their code has problems.
— -
Why am I writing this.
Not to cause panic. Not to pressure specific teams (though if they read this and finally respond to my security reports — great).
I'm writing this because this is a systemic problem nobody talks about loudly enough.
Every day researchers find real vulnerabilities and can't report them properly. Some give up. Some sell to grey market buyers. Some wait for someone less ethical to find the same bug.
And then a $100M hack happens and everyone acts surprised.
This is why it happens.
The fix isn't complicated: separate disclosure from payout. Let anyone report a bug anonymously. Handle identity verification only when actual money changes hands — and even then, make it proportional. On-chain attestation for smaller amounts. Graduated KYC. Safe harbor for pseudonymous researchers who follow responsible disclosure standards.
Until that changes — the people trying to protect this industry will keep being locked out of it.
— -
If you're a protocol team reading this — check your security contacts. Make sure they actually work. Make sure someone reads the emails. The researcher who finds your critical bug before a hacker does might not be able to submit through your preferred platform. That shouldn't mean your users lose money.
If you want to reach me regarding a specific finding — ProtonMail, pseudonym in comments.