More Markets Exploit: How a Liquid Staking Token and E-Mode Pulled $9.3 Million Out of a Lending Market

Around 15.5 million WFLOW drained from the lending market More Markets on August 31, 2026, worth roughly $9.3 million according to the security firm Blockaid. The attacker needed no stolen keys and no gap in the blockchain underneath. They used two building blocks that are wired into almost every larger lending market: a liquid staking token as collateral, and the so-called E-Mode, which treats both sides of a loan as equivalent.

That is precisely why this incident can concern you beyond one small chain. If you have borrowed against staked Ethereum somewhere, or one stablecoin against another, your position is very likely running in the same mode. This article sets out what is established as of today, what remains open, how E-Mode works, and which four details you can look up in your own lending market.

More Markets Exploit: What Happened on Flow EVM on August 31, 2026

More Markets is a non-custodial lending market from More Labs that builds on the Aave V3 codebase and runs on Flow EVM. Users deposit assets there to earn interest, or post them as collateral to borrow against. WFLOW and ankrFLOW are among the supported markets.

According to Blockaid, the incident began on August 31, 2026 at 07:58 UTC. 15.5 million WFLOW disappeared from the reserve labelled mFlowWFLOW. Blockaid explicitly described the figure of roughly $9.3 million as detected impact and not as a final loss figure; the definitive amount is not yet settled, because the transactions are still being traced. The security firm made the incident public first through its channel on the short-message service X, from where several trade outlets picked it up the same day.

More Markets commented briefly on the same day, saying its own team was investigating the reports of an attack and would share its findings. A full post-mortem of the incident is not available at the time of writing. Everything this article says about the sequence of events therefore comes from the security firm's observation and not from the protocol's own analysis.

ankrFLOW and E-Mode: How the Attack Worked According to Blockaid

Blockaid's brief description is that the attacker used a bonded liquid staking token from Ankr together with E-Mode to empty the WFLOW reserve. Put at greater length: the value of the deposited ankrFLOW holdings was set higher in the protocol than it actually was. Against that overvalued collateral, the attacker borrowed real WFLOW and cleared out the reserve with it.

What matters just as much is what that description does not say. Blockaid did not describe either Ankr itself or the Flow blockchain as compromised. On this account, only the More Markets application running on Flow EVM was affected. Whether the weakness sat in the More Markets implementation, in the way the Ankr asset was handled, in the pricing assumptions, or in the interplay between those parts, has not been established so far.

Why a Loan Can Arise Without Price Manipulation

In the two incidents of the past week, the lever lay in the price of a thinly traded token each time. That can be observed and read off the price chart after the fact, as our analysis of the Moonwell exploit on Base on August 30 describes. The route sketched out here is a different one: it does not necessarily require a market price to be driven upwards. It is enough for a protocol to derive the value of collateral from a rule that, under certain conditions, no longer matches reality.

E-Mode Explained: Why a Lending Market Treats Two Assets as Equivalent

E-Mode, written out as Efficiency Mode, is a setting in Aave V3 that permits considerably higher borrowing limits for closely correlated assets. The idea behind it is obvious enough. Anyone posting Ethereum as collateral and borrowing Ethereum carries almost no price risk between the two sides, because it is the same good. Anyone posting staked Ethereum and borrowing Ethereum carries almost no price risk either, because both values normally move in lockstep.

Aave turns this into categories of its own. Each category sets its own values for the assets it contains: the borrowing limit, the threshold at which liquidation kicks in, and the bonus a liquidator receives. The difference is substantial. According to Aave, ordinary borrowing against Ethereum permits around 80 percent of the deposited value, while E-Mode with staked Ethereum as collateral and Ethereum as the loan allows up to 93 percent.

Those thirteen percentage points sound unremarkable, but they change the arithmetic fundamentally. At 80 percent, a fifth of the collateral value remains as a buffer. At 93 percent, seven percent is left. A price drop that would pass without consequence in ordinary mode leads to liquidation in E-Mode. Aave describes the setting in its own documentation on Efficiency Mode and names the underlying assumption there as well: the mode assumes that correlated assets stay correlated.

Liquid Staking Tokens as Collateral: What Sets a Bonded LST Apart From Its Underlying Asset

A liquid staking token is a tradable receipt for a deposited staking position. Anyone staking a cryptocurrency locks it up for a certain period. A liquid staking provider accepts the deposit, takes over the staking, and issues a token in return that can still be traded, lent and posted as collateral. With Ethereum, stETH and rETH are the best-known examples; with Flow it is the ankrFLOW involved here.

The decisive point for any lending market is this: that token is not the same thing as the underlying asset. Its value is derived from a position that can only be unwound after a waiting period. How a protocol sets that derived value is a decision each protocol makes for itself. Some query a market price. Others calculate the value from the ratio of deposited quantity to issued receipts. Both have advantages and drawbacks, and both can come under pressure.

Anyone interested in the yield side of these products will find the providers and their terms in our overview of the best staking platforms. For this article the other side counts: a liquid staking token serving as collateral ties together two risks that were previously separate. The staking risk and the lending market risk then hang on the same position.

What Is Established and What Remains Open in the More Markets Exploit

As of the afternoon of August 31, 2026, the incident breaks cleanly into three parts. The on-chain movement is established: 15.5 million WFLOW left the reserve, and Blockaid named both the triggering transaction and the onward transfers of the funds that followed. The security firm's assessment is likewise established, namely that a bonded liquid staking token and E-Mode together opened the route.

The figure of roughly $9.3 million is flagged as an estimate. Blockaid marked it as detected impact, which means the sum may come out above or below the final number, depending on where the funds went and how much of that can be recovered.

The cause remains open. Nobody has yet evidenced whether the fault lay in More Markets' adaptation of the Aave code, in the parameters of the E-Mode category, in the price source for ankrFLOW, or in some combination of these. Anyone naming an unambiguous cause today is going beyond what is publicly known. That restraint is more than a formality: in the incidents of recent weeks, the first explanation offered has shifted several times once the post-mortem arrived.

Aave V3 Fork Instead of Aave: Why an Offshoot's Parameters Can Differ

A large share of today's lending markets are copies of an established protocol placed on a different chain and fitted out with their own values, rather than independent designs. The technical term for that is a fork. More Markets is one such offshoot of Aave V3.

For you as a user, that produces a difference which is barely visible in the interface. The code may be the same; the numbers are not. Borrowing limits, liquidation thresholds, caps on the borrowable quantity and the choice of price source are set by each offshoot itself, and it does so for a market that is often considerably thinner than the original's. The same setting that is defensible on a deep market can be dangerous on a shallow one.

On top of that comes the question of who is allowed to adjust those values at all, and how quickly that works. The Ajna incident of August 29 showed the opposite pole: there the protocol was immutable and had no governance, which is why there was no pause button. Almost all lending markets sit somewhere between those two ends, and where exactly a protocol stands determines what is possible at all in an emergency.

Checking Your E-Mode Position: Four Details You Can Find in Your Lending Market

This is where the incident turns practical. If you have an open position in a lending market, there are four things you can look up today, and you need neither programming knowledge nor special tools for it. All four appear in the interface of the protocol concerned or in its documentation.

First: is your position running in E-Mode? The setting is usually a toggle in the account view and carries labels there such as E-Mode, Efficiency Mode or Correlated Assets. If it is active, the higher limits of the relevant category apply to you.

Second: how far is your position from the liquidation threshold? Most interfaces show a health factor for this. If it sits close to one, a small movement is enough. The buffer in E-Mode is narrower by construction, so the same numeric value is less reassuring there than in ordinary mode.

Depeg in E-Mode: What Happens When Two Values Drift Apart

A depeg is the drifting apart of two values that are meant to move in lockstep. With a liquid staking token, that means the receipt is worth less on the market than the position it represents. It happens when many holders want to exit at once while unwinding the staking position takes time.

In ordinary mode, drift of that kind is uncomfortable. In E-Mode it can end the position, even though nothing has been lost to you economically. Your collateral still represents the same quantity of the underlying asset, but the price the protocol applies has fallen, and the narrow buffer no longer absorbs it. Aave names exactly that as the principal risk of this setting.

Anyone wanting to see the yield side and the risk side of lending markets next to each other will find the providers and their terms in the comparison of crypto lending platforms; we took apart the underlying mechanics of interest and risk in our article on the interest and risks in crypto lending from August 16, 2026.

Bad Debt: Who Ultimately Carries the Shortfall in a Lending Market

Bad debt describes a loan that no longer has sufficient collateral behind it and can no longer be covered by liquidation either. That gap does not disappear; it travels. In a lending market it hits the depositors of the reserve that was borrowed from first.

On More Markets, the reserve concerned is the WFLOW reserve. Anyone who deposited WFLOW there to earn interest is tied to an event they took no part in and made no decision about. That is the most uncomfortable feature of incidents of this kind, and it repeats: in the incident on Base mentioned above, the open gap hit depositors who had never touched the token that triggered it.

Third in the series of details you can look up: which market exactly is your deposit sitting in? Many protocols separate a core market, in which several assets share a common liability, from isolated markets, in which a shortfall stays contained. This distinction determines whether a shortfall in an entirely different asset can reach you.

Fourth: where does the protocol source the price of your collateral? The answer is in the documentation, usually under headings such as Oracle or Price Feed. If a single trading venue is named there as the source and the asset is thinly traded, you know the weak point.

What You Can Do Without Closing the Position

There is plenty of room between doing nothing and exiting entirely. You can switch E-Mode off if your buffer allows it, and fall back to the ordinary limits. You can add collateral and widen the distance to the threshold. And you can move a position sitting on an offshoot with a thin market over to the protocol whose parameters are carried by broader oversight. Which of these routes makes sense for you depends on your position; none of them is a recommendation for everyone.

Tax and Evidence: What to Document After an Exploit

Regardless of whether this incident affects you, the same advice applies to every shortfall in a lending market: secure the evidence while it is still retrievable. That includes the address of your position, the transaction numbers for the deposit and the outflow, the balance before and after the event, and a dated printout of the protocol interface.

Whether and how a loss of this kind has tax consequences is a question of the individual case and the circumstances, which a tax adviser has to assess. What you can influence yourself is the evidence. Protocol interfaces tend to disappear quickly after incidents, and what you can still download today may be out of reach in a few weeks.

Flow and Ankr: What Is Not Affected According to the Reports So Far

Because reports about incidents often take on a life of their own in circulation, the boundary belongs here explicitly. On Blockaid's account there was no indication that Ankr itself was affected, and none that the Flow blockchain or its infrastructure were impaired. What was described is an incident in a single application running on Flow EVM.

This distinction is no quibble. For you it marks the difference between a chain whose balances are in question and a chain on which one of many applications has taken damage. If you hold assets on Flow that have nothing to do with this lending market, no action is required on what is publicly known so far. How that looks once a full post-mortem is available remains to be seen.

Lending Markets Without Your Own Due Diligence: Where the Responsibility Sits

There is a reason why this kind of incident does not occur in supervised offerings: there a provider makes the decisions about collateral and borrowing, and is liable for them. In an open lending market, that assessment sits with you. It is the price of direct access, and anyone unwilling to pay it will find the supervised alternatives and their terms in the overview of regulated crypto exchanges. For an understanding of the prices such positions run against, a look at our Ethereum price prediction helps, because most E-Mode categories ultimately hang on that value.

The number of incidents in lending markets has been strikingly high in recent days, and all of them followed different routes. Deriving a pattern from that would be premature, because the post-mortems are still outstanding. What can be said is more modest and useful all the same: the building blocks taken apart here are present in many protocols, and their values can be looked up.

Checking E-Mode Collateral: What to Take Away

1. Check whether your position is running in E-Mode and how narrow the buffer really is. The toggle is in your protocol's account view, and the distance to the liquidation threshold is in the health factor beside it. Which providers set which limits is shown in the comparison of crypto lending platforms.
2. Check where the price of your collateral comes from, especially with liquid staking tokens. If the documentation names a single thin trading venue as the source, you know your biggest risk. Anyone wanting to see the yield side of these tokens side by side will find it in the overview of the best staking platforms.
3. Decide deliberately how much due diligence you want to carry out yourself. Anyone unwilling to check an offshoot's parameters regularly is better served by a supervised provider; the terms are in the overview of regulated crypto exchanges.

The incident was picked up by several trade outlets the same day; the fullest account, including Blockaid's figures, is at crypto.news.

(As of August 31, 2026. This article is not investment advice. Prices and fee structures change; check the terms with the provider before you buy.)