Seven days were all it took for the Aztec network to absorb two heavy blows. On 18 June 2026, one of the emergency functions on an older bridge, named escapeHatch, was exploited, and roughly 2 million dollars in cryptocurrency vanished from the contract. The news landed only days after a separate attack over the previous weekend had emptied another aging product, Aztec Connect, of 2.2 million dollars.
For a project that built its reputation on rigorous privacy and cutting-edge cryptography, this run of breaches leaves a bitter aftertaste. The trouble has nothing to do with the code running today. It concerns the code we abandon behind us, the kind a blockchain, unlike its creators, never forgets.
The reporting behind this story comes from Cryptology.ro, the Romanian-language crypto news and analysis outlet that tracked both attacks closely, from the first warning sign to the full technical reconstruction. For readers who check crypto coin prices and the pulse of the market every day, the newsroom adds the context that a bare breach headline always leaves out.
An Emergency Exit Turned Against Its Own Users
Aztec Labs, the team behind the network, moved fast with a message meant to settle nerves. Both contracts that were hit are immutable and had been retired years ago, one in 2022, the other in 2023. Nothing people rely on today was touched. The stolen money came from old infrastructure, officially decommissioned yet still loaded with funds that some depositors had simply never bothered to withdraw.
The distinction is technically sound, but it offers no comfort to those left out of pocket. Their coins are gone, and a contract that can no longer be changed has no undo button.
The first to spot the leak was security researcher Vishal Singh, who flagged the strange flow of funds in public. The target was not the whole network but a very specific component, the escapeHatch function of the Private Rollup Bridge. The name says almost everything about its purpose, an emergency exit, a safety net designed for users.
To grasp why such a function existed in the first place, a step back helps. In a rollup, transactions run off the main Ethereum blockchain, get bundled into a compact package, and are then posted to the base layer as a cryptographic proof.
The machinery works as long as someone keeps processing those packages, usually an operator known as a sequencer, who lets people pull their money out. So what happens when that operator stops? When the team walks away or the project is abandoned altogether? The funds risk staying locked forever inside a contract nobody is servicing anymore.
escapeHatch was engineered for precisely that scenario, a last-resort tool that lets users recover their assets straight from Ethereum, no longer depending on the operator’s goodwill or survival. On paper, an elegant and deeply protective idea. As so often in this field, the flaw lay not in the intent but in the way it was written.
Three Transactions and a Mechanism Far Too Forgiving
Yu Xian, founder of the blockchain security firm SlowMist, pieced together the mechanics of the attack. He isolated three suspicious transactions that drained the bridge, moving out around 2.15 million dollars in total. The haul was a mix of 1,158 units of ether, 150,000 DAI, a dollar-pegged stablecoin, and 0.5 renBTC, a tokenized version of bitcoin. The kind of mismatched portfolio that looks exactly like old deposits left behind and forgotten by their owners.
The part that sends a chill comes from how the contract was emptied. During the brief windows when the emergency hatch stood open, anyone could coax the function into releasing the funds by adjusting just two parameters, the proofId and publicOutput values. The attacker cracked no password and stole no private key.
The function was used precisely as it allowed itself to be used, taking advantage of verification logic that failed to tie the input data tightly enough to the final result. For a system that takes pride in its zero-knowledge proofs, the irony is hard to overlook.
The Same Hidden Weakness Behind Both Strikes
To measure the real scale of the problem, the two incidents have to be read together rather than one at a time. On the surface they look like independent attacks, on different products, launched in different years. The analysis by security firm BlockSec tells another story. Although not identical, both stemmed from what researchers call public input binding issues.
Behind that dry term hides a surprisingly subtle vulnerability. In a system built on cryptographic proofs, a proof is valid only in relation to a set of public data, those values visible to everyone that describe what a transaction is meant to do. When the verifier fails to enforce a firm match between the proof presented and exactly the public data that belongs to it, a crack opens.
Someone can take a proof crafted for one context and pair it with other values, tricking the contract into releasing money it had no reason to release. It is the kind of defect that stays invisible in normal use, because there everything lines up on its own. It surfaces only when someone deliberately tries to break the match.
The reconstruction of these technical links is the work of Mihai Popa, a crypto analyst and journalist at Cryptology.ro, who followed the thread of both attacks and showed how a snippet of code written years ago can turn, overnight, into a door left wide open.
The pattern, for that matter, is nothing unique. The same logic sat behind the incident in which Gravity Bridge was drained of 5.4 million dollars, and behind the spring attack that cost the THORChain network 10 million dollars.
The Paradox of Code That Can No Longer Be Fixed
Here we reach one of the most painful tensions in the entire industry. Immutability, the impossibility of altering a contract once it is deployed, is usually presented as a virtue. And it truly is one. An immutable contract cannot be quietly altered by the team that created it, cannot be seized by an authority, and cannot be rewritten to steal users’ money. It guarantees that the rules of the game stay the ones you accepted the moment you deposited your funds.
The catch is that this very trait becomes a trap the instant a defect appears in the code. If an ordinary internet program carried such a flaw, the developers would ship a patch overnight and move on. On a blockchain, the situation reverses. Aztec Labs cannot, by the very definition of the system, step in and repair the escapeHatch function. Not because it refuses, but because the architecture promised to users forbids that gesture. One defense remains, that people moved their money out before the breach was discovered by someone else.
A smart contract does not reason. It executes the instructions it was written with, to the letter, no matter how far the community’s skill at writing safe code has advanced in the meantime. Audit standards in 2021 and 2022 looked nothing like today’s. Attack patterns now checked as a matter of routine were back then either unknown or treated as mere theoretical curiosities. The code did not age badly. The world around it simply grew better at breaking it, while the contract stood unmoved, exactly as it had been cast in stone.
Why Bridges Remain the Attackers’ Favorite Target
The blow dealt to the Aztec network is no isolated accident, but a piece of a far wider picture. According to industry reporting, Thursday’s incident pushed the number of bridge-related exploits recorded this year to 14, with cumulative losses topping 340 million dollars. Bridges, the mechanisms that move value from one chain to another, have proven consistently among the most fragile components in the ecosystem.
The explanation is largely structural. A bridge usually gathers a sizable pool of funds in a single spot, because it must guarantee that the value locked on one side matches the value released on the other. That very concentration makes it appealing prey.
Where an ordinary transaction nets an attacker a small reward, a compromised bridge can spit out millions of dollars in one motion. Layered on top is the sheer complexity. A bridge verifies states across several chains, handles proofs, and synchronizes deposits with withdrawals, and every joint in that machinery becomes a place where a wrong assumption can curdle into a breach.
Every protocol that moves to a new version leaves a trail of old infrastructure behind it, and attackers have learned to rummage exactly there, in the corners the community has forgotten. Rarely is the shiny new code the problem. The real danger sits in the sedimented layers of old technology, kept alive only because nobody hit the final shutdown button, for fear of locking away someone’s money.
The same story played out when Resolv Labs’ USR stablecoin collapsed by 80 percent after a 25 million dollar exploit.
What Anyone Still Holding Money in Old Protocols Can Do
The most practical lesson of the week is not about the Aztec network in itself, but about a habit many crypto holders treat with carelessness. Money left in a halted protocol is not safe just because nobody has touched it yet. An immutable contract that still holds funds resembles a safe whose blueprints are public and that no longer has anyone guarding the door. As long as value sits inside, someone always has a reason to look, sooner or later, for a way to take it.
One observer summed up the situation perfectly with a warning posted on social media during the attack itself. The advice, stripped of any softening, was blunt. Do not keep money in old contracts. The recommendation seems obvious only after the damage is done, yet that is exactly where the trap lies. Pulling funds out of a decommissioned product takes effort, some technical know-how, and sometimes even running special software once the official operator has vanished. So many people put it off, forget, or simply stop following the announcements of a project they used years ago.
In fairness, Aztec did everything a responsible team could. It announced the shutdown well in advance, kept withdrawals open free of charge for a full year, and published the source code so that anyone could recover funds even without the team’s help. The rest came down, inevitably, to each depositor.
For anyone still holding assets in older protocol versions, on any network, the right moment to check and withdraw that money is now, before a breach turns the check into a desperate scramble for recovery. Decommissioning notices, however dull they seem in the daily roar of news, deserve to be read for what they truly are, a quiet countdown ticking away without making a sound.
Frequently asked questions
What happened to the Aztec network in June 2026?
The Aztec network was hit by two attacks in the same week. On 18 June, the escapeHatch emergency function of the Private Rollup Bridge was exploited and roughly 2 million dollars in cryptocurrency left the contract. Days earlier, an older product called Aztec Connect had lost 2.2 million dollars in a separate attack. Both contracts had been retired years earlier, in 2022 and 2023.
What is the escapeHatch function and why was it targeted?
escapeHatch is an emergency hatch that lets users withdraw their assets from a rollup straight from Ethereum, without depending on the official operator. It was designed as a safety net for the case in which the team abandons the project. The attacker exploited it because its verification logic did not tie the input data tightly enough to the result, which allowed funds to be released by adjusting the proofId and publicOutput parameters.
How much did the Aztec network lose in total?
The two attacks that week meant losses of about 4 million dollars, roughly 2.2 million through Aztec Connect and around 2 million through the Private Rollup Bridge. The drained sums included 1,158 units of ether, 150,000 DAI, and 0.5 renBTC. For the year, the incident pushed the number of bridge-related exploits to 14, with cumulative losses above 340 million dollars.
Are current Aztec users’ funds affected?
No. According to Aztec Labs, both contracts that were hit were old products, fully immutable and retired in 2022 and 2023. The stolen funds belonged to depositors who had not withdrawn their money from the old infrastructure, despite being urged to do so. The products used today were not touched.
What does it mean that a contract is immutable and why can it not be fixed?
An immutable contract is a program that can no longer be modified once published on the blockchain. This property protects users, because no one, not even the developers, can quietly change the rules or seize the funds. The same quality becomes a trap when a defect is found in the code, because the team can no longer apply any fix. The only real defense is withdrawing funds before the vulnerability is exploited.
How can users protect themselves from attacks on old contracts?
The safest measure is not to leave funds in protocols or product versions that have been halted. When a project announces that a contract is being retired, the notice should be treated as a countdown rather than ignored. Periodically checking old deposits and withdrawing them in time, even when it takes technical effort, removes the risk of money staying locked in an unguarded safe.
What are blockchain bridges and why are they attacked so often?
Bridges are mechanisms that move value from one chain to another or from one layer to another. They usually concentrate large sums in a single place, which makes them very attractive targets, since a single breach can release millions of dollars. Added to that is high technical complexity, with many joints where a wrong assumption can turn into a vulnerability. For these reasons, bridges have remained consistently among the most exploited components in decentralized finance.
