Loss Exceeds $7.5 Million: Analysis of the Honeypot Attack Targeting MEV Bot and Tracking of Stolen Funds

marsbitPublished on 2026-06-22Last updated on 2026-06-22

Abstract

On June 21, Jaredfromsubway.eth, one of Ethereum's most active MEV bots, suffered a sophisticated honeypot attack, losing over $7.5 million in crypto assets. Attackers deployed a complex contract system, including a coordinator, trigger, and fake token/pair contracts, to exploit the bot's arbitrage logic. The core attack trick involved baiting the bot with transactions: small trades appeared normal and consumed token approvals, while larger trades preserved these approvals by using fake swaps. This left the bot with large, unused approvals for assets like USDC, USDT, and WETH. The attacker then drained these approvals in a final step. The stolen funds—$2.87M USDC, $2.04M USDT, and 1,474 WETH—were consolidated and partially laundered through Tornado Cash. This incident highlights that MEV bots must not rely solely on simulated profits for security. Extra caution is required with unfamiliar contracts or custom wrappers in arbitrage paths, and post-transaction allowance checks are strongly recommended.

On June 21, Jaredfromsubway.eth, one of the most active MEV Bots on the Ethereum network, fell victim to a meticulously orchestrated "honeypot attack," resulting in the loss of over $7.5 million worth of crypto assets. The following is Beosin Security Team's analysis of this attack and tracking of the stolen fund flow.

Attack Process Analysis

Attack Contract Family

- Coordinator Contract (0xb84db016324e8f2bfdd8dd9c260338aee0a8df52): Responsible for recording whether the current block is in an 'armed' state and, in the final stage, cyclically calling sub-contracts to drain funds.
- Trigger Contract (0x4de8c729a064ff6087cc84a4152969349e4feb98): Responsible for setting the state of fake trading pairs within the same block, making the arbitrage path appear executable.
- Sub-contract / Fake Token Contract: Disguised as a normal ERC-20 token, used to obtain genuine approvals.
- Hub Contract: Responsible for paying out a small amount of real profit, making the MEV Bot believe the opportunity is profitable.
- Ring V2 Pair: Fake Uniswap v2 trading pair.
- Fake Intermediate Token Contract: Used to construct multi-hop arbitrage paths, such as fCAP, fUSDC.

The Key to the Attack: Deceiving Approvals

By analyzing on-chain transactions, the attacker constructed multiple sets of bait transactions:

- Large USDC: The bot profited about 36.997120 USDC, but left behind an approval for 20 USDC.
- Large USDT: The bot profited about 37.053440 USDT, but left behind an approval for 20 USDT.
- Large WETH: The bot profited about 0.0179 WETH, but left behind an approval for 16 WETH.
- Small transactions behaved normally, with the approval consumed within the same transaction to reduce suspicion.

In small transactions, after the bot authorized a real token allowance, the sub-contract would immediately transfer the real tokens away. The approval was consumed, appearing completely normal.

In large transactions, however, the sub-contract would not call `transferFrom` to move the real tokens. Instead, it would directly mint fake tokens via the fake trading pair. The bot thought it had completed the normal pre-swap steps, but the real token approval remained intact.

This is the core of the entire attack: small transactions normally consume approvals, while large transactions preserve them.

Attack Procedure

Taking the attack transaction targeting USDC as an example:

(1) The attacker calls the coordinator to set the current block as 'armed'.
(2) The attacker calls the trigger to update the state of multiple fake Ring V2 pairs.
(3) The MEV Bot detects the arbitrage opportunity and executes the transaction.

The internal process of the MEV Bot transaction is roughly as follows:

(1) The MEV Bot contract approves a large USDC allowance to a specific sub-contract.
(2) The MEV Bot calls the `wrapTo`/`wrap` function of the sub-contract.
(3) Because the current state is 'armed', the sub-contract does not consume real USDC. Instead, it mints fake tokens to the pair, and the USDC approval is preserved.
(4) The MEV Bot continues by calling the `swap` function on the fake pair.
(5) The second-hop pair sends the tokens to the MEV Bot.
(6) The hub contract pays a small amount of real USDC profit to the MEV Bot.

approval example

tx hash: 0x0121e07a916c06eea3e7daf11893f3f0b95b9e1684124545ae14c32aee6029bb

The result seen by the MEV Bot: A successful arbitrage transaction yielding real USDC profit. However, the USDC approval was retained by the sub-contract.
This process was repeated for USDC, USDT, and WETH, ultimately accumulating numerous approvals.

The attack transaction hash is:

0x2be8704f5a59b69e0b71f64aefdb99eb0e8ae9fb3926147c581910d71bcf3e65

The attacker calls the `drain loop` function of the coordinator contract. The calldata contains 66 sub-contract addresses and the MEV Bot contract address. For any sub-contract to which the MEV Bot contract had previously left an allowance, that sub-contract could directly transfer the corresponding real tokens to the attacker.

Final Result:

- The entire 20 USDC large allowance was drained.
- The entire 16 WETH large allowance was drained.
- Part of the USDT allowance still exists, but the USDT balance is insufficient.

Fund Flow Analysis

After succeeding, the attacker's address (0x3e37f4A10d771Ba9dE44b6d301410b1BEdeA65d0) received $2.87M USDC, $2.04M USDT, and 1,474 WETH. Subsequently, the attacker exchanged the stablecoins for ETH and transferred it to the following 4 addresses:

- 0xe3Da36E4bd1a5738fa5D6Ef4F0e4dF40bDeB5f17 (approx. 1,000 ETH)
- 0x74Dc5b93586D248D5Aec64b3586736FF0A0D0e65 (1,001 ETH)
- 0xd8C125efCBc99408eC8723E9BBd81d1E8D39D845 (1,001 ETH)
- 0x71d4416A7A85e08a5Fe7227Ca3B44Fc639e94e97 (1,423 ETH)

Among these, 0xe3Da3 has transferred 1,000 ETH to Tornado Cash. The ETH at the other three addresses has not seen further movement. The fund flow chart is shown below:

Conclusion

This attack demonstrates a highly sophisticated method: instead of directly attacking contract code, the attacker, based on the business logic of MEV Bots, constructed corresponding arbitrage scenarios to mislead the MEV Bot into making seemingly harmless approvals, and then transferred its assets. For arbitrage bots and MEV Bots, relying solely on simulated profits to judge the safety of a route is insufficient. Especially when an arbitrage path involves unfamiliar contracts, fake tokens, or custom wrappers, caution is essential. Consider enforcing checks on allowance changes after transactions.

View Original Article

Trending Cryptos

ctr
CitreaCTR
wstusdt
wrapped stUSDTWSTUSDT
apr
aPrioriAPR
ctx
Cryptex FinanceCTX
audio
AudiusAUDIO
comp
CompoundCOMP
mantra
MantraMANTRA
xvs
VenusXVS
waxl
AxelarWAXL
bill
Billions NetworkBILL
pyth
PYTH (Pyth)PYTH
rune
THORChainRUNE
velodrome
Velodrome FinanceVELODROME
brev
BrevisBREV
zrx
ZRX（0X）ZRX

Related Questions

QWhat was the total loss suffered by the MEV bot Jaredfromsubway.eth in the honeypot attack on June 21st?

AThe total loss exceeded 7.5 million US dollars worth of crypto assets.

QWhat was the core mechanism used in the honeypot attack to exploit the MEV bot?

AThe core mechanism was deceiving the bot into granting excessive token approvals (allowances). Small, normal-looking trades consumed the approval, while larger trades left the approval intact. The attacker later exploited these retained approvals to drain the bot's funds.

QWhich specific real-world assets (tokens) did the MEV bot lose in this attack?

AThe bot lost USDC, USDT, and WETH. Specifically, it lost $2.87M USDC, $2.04M USDT, and 1,474 WETH (which was later converted to more ETH).

QWhat was one of the final destination addresses for the stolen ETH, as mentioned in the fund flow analysis?

AOne of the final destinations was the privacy mixer Tornado Cash, where 1,000 ETH from address 0xe3Da36E4bd1a5738fa5D6Ef4F0e4dF40bDeB5f17 was sent.

QWhat key security recommendation does the article provide for MEV bots and arbitrage robots to prevent similar attacks?

AThe article recommends that MEV bots should not rely solely on simulated profits to judge the safety of a route. They should be cautious with unfamiliar contracts, fake tokens, or custom wrappers in an arbitrage path. Furthermore, they should consider implementing mandatory checks for allowance changes after a transaction.

Related Reads

As Channel Dividends Fade, What Can DeFi Protocols Rely On to Resist Giant Harvesting?

The article discusses how major tech companies like Coinbase, Stripe, and Kraken are increasingly acquiring or building their own underlying infrastructure to capture profits, moving away from relying on external DeFi protocols. It uses the example of Coinbase's Base blockchain, which generates sequencing fees for Coinbase from protocols like Morpho that deploy on it. Similarly, Stripe acquired Bridge to capture stablecoin yield instead of paying it to Circle, and Kraken bought NinjaTrader for its derivatives licenses. The core argument is that when channel-based advantages (like user traffic) fade, control over the profit-generating infrastructure becomes critical. While this could lead to a future dominated by a few integrated giants, the article suggests a more balanced outcome is possible. Protocols that have already achieved deep, multi-chain integration and embedded themselves into corporate backends create high switching costs, making them difficult to replace. Examples include Morpho (despite reliance on Base, it's widely deployed) and Uniswap. The piece concludes that the race between institutional expansion and open-source protocol proliferation will ultimately shape the industry's structure.

marsbit3m ago

As Channel Dividends Fade, What Can DeFi Protocols Rely On to Resist Giant Harvesting?

marsbit3m ago

The Computing Power Dilemma in the Sino-US AI Rivalry

The Sino-US AI rivalry faces a fundamental bottleneck: the widening compute power gap. While Chinese AI chip companies have seen investment surges, their current focus remains largely on the less demanding inference market. The real challenge lies in the high-end training chip sector, crucial for developing cutting-edge large language models (LLMs), where Nvidia holds a near-monopoly. The compute disparity is stark. US tech giants like Meta, Google, and xAI command massive GPU clusters, enabling them to train trillion-parameter models rapidly. Estimates suggest US data center count and total compute capacity significantly outstrip China's. This "brute force" advantage allows for faster model iteration and exploration of larger parameter scales, with top US models reportedly leading their Chinese counterparts by 8 to 15 months. Chinese alternatives, such as Huawei's Ascend and others from companies like Moore Thread and Biren, are emerging. They show promise in inference and some training scenarios, closing the performance gap with mid-range Nvidia products. However, the core hurdle extends beyond raw chip performance to the entrenched software ecosystem, exemplified by Nvidia's CUDA platform. The path forward involves "walking on two legs": navigating import restrictions while heavily investing in the domestic chip industry. Though still in a catch-up phase, China's vast market, talent pool, and capital are fostering progress. The ultimate test is whether Chinese firms can build a competitive hardware-software ecosystem to power the next generation of AI.

marsbit3m ago

The Computing Power Dilemma in the Sino-US AI Rivalry

marsbit3m ago

He Kaiming's Team's New Work: After Deleting VAE and Private Data, Text-to-Image Generation Becomes Even Stronger

KaiMing He's team introduces **MiniT2I**, a minimalist text-to-image (T2I) model that challenges the complexity of mainstream approaches. It eliminates components commonly considered essential: the VAE encoder-decoder, AdaLN conditioning mechanisms, auxiliary losses, private training data, and post-training alignment stages like RL/DPO. Instead, it uses a pure flow-matching objective trained directly on RGB pixels. The model employs a simplified **MM-JiT** Transformer architecture. It removes AdaLN blocks for conditioning and instead prepends two lightweight text adapter blocks to a standard pre-norm Transformer, allowing frozen T5 text features to adapt to the denoiser. Training follows a two-stage, LLM-like paradigm using only public datasets: pre-training on LLaVA-recaptioned CC12M for coverage, followed by fine-tuning on ~120k high-quality image-text pairs. With just 258M parameters (B/16), MiniT2I achieves competitive scores (0.87 on GenEval, 84.2 on DPG-Bench), outperforming larger pixel-space models. Scaling to 912M parameters (L/16) yields results comparable to SD3-Medium (~2B parameters) in style, composition, and imagination, though it lags in text rendering and named entities due to public data limitations. Key advantages include lower computational cost (~570 GFLOPs vs. ~1379 for latent models) and architectural simplicity. Acknowledged limitations include patch boundary artifacts in pixel space, side effects of high CFG scales, resolution ceilings for sequences longer than 1024 tokens, and the aforementioned data bottlenecks. The work demonstrates that high-performance T2I generation is possible with a radically simplified, publicly reproducible baseline.

marsbit8m ago

He Kaiming's Team's New Work: After Deleting VAE and Private Data, Text-to-Image Generation Becomes Even Stronger

marsbit8m ago

The Insurance Industry Faces Its Biggest Competitor: Are Prediction Markets the "Barbarians at the Gate"?

The insurance industry, long a stable "ballast" in the economy, may face a significant challenge from the rise of prediction markets, which are beginning to function as a new form of risk hedging and insurance. Platforms like Kalshi and Polymarket are demonstrating their utility in areas traditionally dominated by insurers. Examples include Kalshi's partnership with sports insurance broker Game Point Capital to offer more cost-effective hedging for NBA team performance bonuses, and Polymarket's collaboration with real estate platform Parcl, allowing users to hedge against housing price fluctuations in major US cities. A New York bar also used Kalshi to hedge a marketing promotion tied to an NBA game outcome, highlighting prediction markets' potential for small business risk management. These markets offer advantages over traditional insurance and sports betting in transparency, liquidity, and flexibility. They allow information monetization across a wider range of events, act as neutral platforms rather than direct counterparties, and provide clearer pricing. A historical precedent is the "Mattress Mack" marketing campaigns, which used sports betting for large-scale customer refunds, but prediction markets offer a more systematic and accessible model. Experts like SIG CEO Jeff Yass see their potential for efficient, parameter-based risk sharing, such as for weather-related property damage. However, challenges remain, including liquidity issues, unclear regulatory boundaries, and potential manipulation of event outcomes. Despite these hurdles, prediction markets represent a growing competitive force for both traditional gambling platforms and segments of the insurance industry.

marsbit9m ago

The Insurance Industry Faces Its Biggest Competitor: Are Prediction Markets the "Barbarians at the Gate"?

marsbit9m ago

Insurance Industry Faces Its Biggest Competitor: Are Prediction Markets the 'Barbarians at the Gate'?

"Insurance Industry Faces New Rival: Are Prediction Markets the 'Barbarians at the Gate'?" Prediction markets, exemplified by platforms like Kalshi and Polymarket, are emerging as potential disruptors to the traditional insurance industry by offering alternative risk-hedging mechanisms. These markets allow users to bet on specific event outcomes, effectively creating a form of customizable, on-demand insurance. Key examples highlight this shift. In sports, Kalshi partnered with insurance broker Game Point Capital to provide NBA teams with more affordable options to hedge performance bonuses compared to traditional insurers. In real estate, Polymarket's collaboration with Parcl lets users speculate on city-specific housing price indices, allowing homeowners to hedge against price drops or buyers against price increases. Furthermore, businesses like a New York bar have used Kalshi to hedge marketing promotions (e.g., offering free drinks if a team wins), framing the transaction explicitly as placing a "hedge." The article argues prediction markets offer advantages over traditional insurance and even sports betting in transparency, liquidity, and flexibility. They provide a wider range of event coverage, act as neutral platforms rather than counterparties, and offer clearer pricing. The piece cites historical precedents like large "refund promotion" hedges by businesses using sportsbooks but notes prediction markets modernize the concept. However, challenges remain for widespread adoption as an insurance alternative, including limited liquidity in some markets, unclear regulatory status, and potential vulnerabilities in event resolution mechanisms. Despite these hurdles, prediction markets are positioning themselves as new tools for risk management, directly challenging certain segments of the conventional insurance landscape.

Odaily星球日报15m ago

Insurance Industry Faces Its Biggest Competitor: Are Prediction Markets the 'Barbarians at the Gate'?

Odaily星球日报15m ago

Trading

Spot
Futures

Hot Articles

Discussions

Welcome to the HTX Community. Here, you can stay informed about the latest platform developments and gain access to professional market insights. Users' opinions on the price of ETH (ETH) are presented below.

活动图片

Top Questions

Hot Categoriesright-arrow

Hot Tagsright-arrow