Billions are traded on decentralized exchanges every day, with the latest figures from DefiLlama showing that DEX volume has reached around $5 billion in the last 24 hours. Aggregators assess the rates and routes available from different liquidity providers to find the best deals. To illustrate, while 0x collects data from about 150 services, Uniswap’s quote API shows the user both the price and the route.
A number of services also offer a so-called transaction simulation, which assumes that the results correspond to what would happen in reality. New research from DeFi infrastructure firm Enso, published on July 16, 2026, reports that trust can be manipulated – and the company has identified pools that do just that.
The risk extends beyond these two contracts. Wallets, DEX aggregators and other consumer-facing apps often rely on off-chain simulations to find users the best trade. Enso discovered that a pool can exploit this process by showing an attractive price to win the route during the simulation, and then offering a worse price when the actual transaction reaches the blockchain.
Enso calls these setups ‘poisonous pools’. They benefit from the gap between a simulated transaction and what actually happens in the chain. By checking values like tx.gasprice, tx.origin and block.coinbase, a smart contract can tell when a transaction is only simulated and behave normally during the preview, and then change its behavior when the real transaction takes place.
Two real cases, two types of damage
After two months of analyzing RPC data, transaction traces and contract data with the support of Curve Finance and Oku contacts, Enso managed to identify pools functioning on Ethereum and Polygon.
On Ethereum, a curve $USDC/USDT-pool used a manipulated price oracle to make simulated trades look better than the real thing. During a simulation, the oracle applied a reduced rate. But when actual trading continued, that discount disappeared, leaving users with less than expected. The manipulation was also difficult to spot because the oracle relied on legitimate Chainlink price feeds and only changed their output when it discovered a simulation.
Enso calculated that the pool quotes were inflated by almost $225,000, although this does not mean that $225,000 was stolen. The operator’s recorded net profit was $34,592.87, with nearly $23,440 being Curve’s normal fees. Enso also discovered 129,070 swaps where users received quotes lower than what they should receive, and saw 37,425 failed transactions that are still costing their users gas.
The Polygon case caused more disruption than profit. A $USDC/WETH pool using a Uniswap v4 hook charged approximately 98.9% fee when gas prices rose above 100 gwei and used other signals to detect simulations. Although the pool made little money, it was repeatedly shown to offer the best route during simulations, only for the actual transactions to fail. That wasted gas and other resources through routers and trading systems.
Enso recorded 37,467 failed swaps, bringing the failure rate to 99.1%. About 93% of these reversed trades involved MEV and arbitrage bots. One cross-DEX operation alone reportedly went through approximately 2,525 burner wallets and yielded 3,509 failed trades.
A new look at DeFi security
The Curve pool did not behave maliciously all the time, making it harder to detect. Enso found that the discount setting changed 26 times in 48 observed periods, with the pool exhibiting malicious behavior during approximately 59% of the monitored hours. In other words, a pool that looks secure on one check can become hostile later, making static code reviews and reputation filters less reliable.
Enso also found signs that the approach may be repeated. According to the report, the same operator used EIP-7702 smart accounts linked to a shared deployment to deploy 18 oracle contracts behind at least six similar pools.
The impact was not limited to bots and professional traders. Enso says MetaMask routed 6,625 swaps through the toxic Curve pool, showing how flaws in the routing infrastructure can also affect regular users.
Enso does not blame Curve, Uniswap or the wallets that communicated with these pools. Their systems, the company says, are built around a reasonable assumption: that a liquidity pool will behave the same way during a simulation as it does when the actual transaction takes place.
“The sector has been busy optimizing price developments for years,” says Milos Costantini, co-founder and chief product officer of Enso. He added: “Our findings suggest that the next challenge is to verify the integrity of the implementation.”
Both identified locations became inactive after Enso’s disclosure. The Polygon hook has reportedly been inactive since May 15, while the Curve pool’s discount is now at zero. Enso says it has added toxic pool detection to Enso Shield, using live on-chain context and historical quote behavior instead of relying on a single simulation.