The researcher who ran an AI crypto security audit on Zcash — and uncovered a flaw serious enough to crater the coin’s price by 38% — has confirmed his next target is Monero. Taylor Hornby, a security engineer hired by nonprofit developer Shielded Labs, told followers on X that he’ll be adding XMR to his audit queue after the dust settles on what turned out to be one of the most consequential cryptocurrency vulnerability disclosures in recent memory.
- An AI crypto security audit by researcher Taylor Hornby exposed a four-year-old Zcash flaw that sent ZEC crashing 38%.
- Hornby confirmed Monero is next on his AI crypto security audit list, raising fresh concerns across the privacy coin sector.
- The Zcash Orchard bug, undetected since 2022, could have allowed unlimited undetectable counterfeit ZEC to be minted by attackers.
- Shielded Labs hired Hornby in April to find protocol flaws; he disclosed the bug rather than exploit it and plans to seek a coinholder grant.
- An AI crypto security audit by researcher Taylor Hornby exposed a four-year-old Zcash flaw that sent ZEC crashing 38%.
- Hornby confirmed Monero is next on his AI crypto security audit list, raising fresh concerns across the privacy coin sector.
- The Zcash Orchard bug, undetected since 2022, could have allowed unlimited undetectable counterfeit ZEC to be minted by attackers.
- Shielded Labs hired Hornby in April to find protocol flaws; he disclosed the bug rather than exploit it and plans to seek a coinholder grant.
Table of Contents
A Four-Year-Old Flaw Hidden in Plain Sight
To appreciate why Hornby’s Monero announcement matters, you need to understand what he actually found in Zcash. The bug sat inside the blockchain’s Orchard privacy pool — a component introduced in 2022 — and had been sitting there, completely undetected, for roughly three years. What made it dangerous wasn’t just that it existed, but what an attacker could have done with it: mint an unlimited number of counterfeit ZEC tokens without leaving any trace detectable on-chain.
Think about that for a moment. A bad actor with knowledge of this flaw could theoretically have been inflating the ZEC supply in secret, draining real value from every legitimate holder, with no forensic breadcrumb to follow. Hornby discovered it on May 29. Shielded Labs disclosed it publicly on Thursday and pushed an emergency patch through by June 1 — a remarkably fast turnaround for a protocol-level fix, but the market didn’t wait for reassurances. ZEC dropped 38% in the 24 hours that followed.
The price collapse wasn’t purely rational panic — it reflected a real and uncomfortable question the community couldn’t immediately answer: had anyone already exploited this? Because the Orchard pool is shielded by design, transactions inside it are opaque. If someone had been quietly minting counterfeit ZEC for months or years, the historical record wouldn’t necessarily show it. That ambiguity is precisely what makes privacy coin bugs so uniquely unsettling compared to vulnerabilities in, say, a conventional smart contract on Ethereum. It also explains why an AI crypto security audit capable of surfacing such hidden flaws is becoming an essential tool rather than an optional extra.
How an AI Crypto Security Audit Found What Humans Missed
The method Hornby used is what’s drawing as much attention as the finding itself. He conducted his AI crypto security audit using Anthropic’s Opus 4.8 model, one of the most capable reasoning-focused large language models currently available. Hornby was brought on by Shielded Labs in April specifically to hunt for protocol-level bugs — the kind of deep cryptographic and logic flaws that can slip through even rigorous human code review, especially when a codebase grows large and complex over years of iterative development.
The fact that Opus 4.8 helped surface a bug that had evaded detection since 2022 raises a legitimate question for every other privacy protocol team out there: if your codebase hasn’t been put through a similar AI-assisted review, what assumptions are you making about its safety? Traditional audits, even thorough ones by experienced cryptographers, operate at human reading speed. AI models can pattern-match across thousands of lines of code at a pace no human team can match, and they don’t get fatigued or skip sections they assume are already correct.
That’s not to say an AI crypto security audit is foolproof — a model can miss things too, and it can produce false positives that waste engineering time. But the Zcash episode is already being cited in security circles as a proof-of-concept for AI-assisted protocol auditing that’s hard to dismiss. It found a real, critical, exploitable flaw. That’s the benchmark.
Why Monero Is a More Complex Target
When someone on X asked Hornby whether he’d turn his attention to other privacy coins, his response was direct: “Absolutely! I’ll add Monero to my queue of things to audit.” It’s a casual reply that carries significant weight for anyone following the XMR ecosystem.
Monero and Zcash both aim to give users financial privacy, but they take meaningfully different approaches. Zcash gives users a choice — you can transact with a transparent address (visible on-chain, similar to Bitcoin) or a shielded one (encrypted via zero-knowledge proofs). Most ZEC transactions, historically, have actually been transparent. Monero offers no such choice: privacy is mandatory. Every transaction hides the sender, receiver, and amount by default, using a combination of ring signatures, stealth addresses, and RingCT (Ring Confidential Transactions).
That architectural difference has a direct implication for security auditing. With Zcash’s transparent layer, at least some transaction history is inspectable. With Monero, there’s no such fallback. A supply inflation bug in XMR’s protocol — analogous to what Hornby found in Zcash — would be essentially invisible from the outside. No one watching the blockchain could tell whether the circulating supply matched what it should be. It’s worth remembering that Monero’s total supply is already harder to verify independently than Bitcoin’s, and that opacity is a feature by design. But it also means the stakes of an undiscovered bug are arguably higher. Running a rigorous AI crypto security audit against Monero’s codebase is therefore a considerably more demanding undertaking than the Zcash exercise — and arguably a more urgent one.
The Broader Warning for the Crypto Security Landscape
Hornby’s decision to disclose the Zcash flaw rather than exploit it was, by his own account, personal as much as professional. He described the Zcash developer community as being “like family” and said he couldn’t “live with that kind of betrayal.” That’s a reassuring sentiment — but it also illustrates how much of crypto security still depends on the ethics of individual researchers rather than systematic safeguards.
What happens when the person holding that knowledge isn’t Hornby? The bug existed for three years. Theoretically, someone else could have found it first and chosen a very different path. The crypto industry has a long and uncomfortable history of exactly that scenario playing out — from the $600 million Poly Network hack in 2021 to the $320 million Wormhole bridge exploit in 2022. Those were smart contract vulnerabilities, but the pattern is the same: a flaw exists, someone finds it, the outcome depends entirely on who that someone is. Conducting a proactive AI crypto security audit before a malicious actor gets there first is precisely the shift in approach the industry needs.
For his part, Hornby plans to apply for a Zcash coinholder grant to fund his continued research — a reasonable model for a space where protocol security is a public good but doesn’t always attract consistent funding. Bug bounty programmes exist across the industry, but they vary wildly in scope and payout. A dedicated researcher running systematic AI-assisted audits across multiple protocols is a different proposition entirely, and it’s not obvious who pays for that at scale.
AI Crypto Security Audit Work Could Reshape How Protocols Defend Themselves
The Zcash incident is likely to accelerate conversations that were already happening quietly among protocol teams about integrating AI tools into their security processes. The barrier to entry has dropped significantly — Anthropic’s models are accessible via API, and while running a thorough AI crypto security audit still requires someone with deep cryptographic knowledge to interpret and validate what the model surfaces, the combination of human expertise and AI assistance is clearly more effective than either alone.
Whether Hornby’s Monero audit finds anything significant remains to be seen. XMR’s codebase has been reviewed many times by independent researchers, and the Monero Research Lab maintains active cryptographic review. But the Zcash experience is a reminder that confidence built on past audits has an expiry date. Code evolves, new components get added, and assumptions made in one era of cryptographic thinking can become liabilities in another.
The more interesting long-term question might not be what Hornby finds in Monero, but whether his approach — a systematic, AI-assisted AI crypto security audit across the privacy coin ecosystem — becomes a template that other researchers and protocol teams adopt before the next flaw surfaces on its own terms. Given what just happened to ZEC, the incentive to find out first has never been clearer.
Source: CoinDesk
Frequently Asked Questions
What is an AI crypto security audit and how did it find the Zcash bug?
An AI crypto security audit uses large language models to analyse protocol code for logic flaws faster than traditional manual review. Researcher Taylor Hornby used Anthropic’s Opus 4.8 model to find a bug in Zcash’s Orchard privacy pool that had gone undetected since May 2022, which could have enabled unlimited counterfeit ZEC minting.
What exactly was the Zcash Orchard bug?
The flaw lived inside Zcash’s Orchard privacy pool and had been dormant since May 2022. If exploited, it could have let an attacker mint an unlimited number of counterfeit ZEC tokens entirely undetected, bypassing the transparency checks that would normally catch such inflation.
Why is Monero considered a higher-risk target than Zcash?
Unlike Zcash, where users can choose between transparent and shielded addresses, Monero hides transaction details by default for every user. That means any undetected inflation bug in Monero’s protocol would be significantly harder to spot after the fact, making a proactive audit especially important.
Who is funding Taylor Hornby’s privacy coin audit work?
Shielded Labs, a nonprofit developer working on the Zcash network, hired Hornby in April. For future work, including the planned Monero audit, Hornby says he intends to apply for a Zcash coinholder grant to cover costs.
