AQ FORENSICS

BLOCKCHAIN INTELLIGENCE

+43 676 513 574 8OFFICE@AQ-FORENSICS.COM

BLOG

Monero on Hyperliquid – Blockchain Forensics with PowerShell Skript
AQ INSIGHTS

Monero on Hyperliquid: How We Continued Tracing the Assets with PowerShell Skript after ETH and USDC Conversions

The Initial Situation

The investigation began with the theft of ETH. The subsequent movements could initially be traced using professional on-chain analysis methods. The analysis showed that the perpetrators first converted the stolen ETH into USDC and then into XMR1, a representation of Monero used on Hyperliquid.

The reconstructed asset path initially followed this sequence:

  • ETH → USDC → XMR1 → Hyperliquid

One transfer central to the subsequent investigation involved exactly 70,500615 XMR1. Hyperliquid's ledger data assigned a value of 35.151,606639 USDC to this transaction at the time. Up to this point, the flow of funds could be traced using professional on-chain analysis tools.

Then the situation changed.

Both the professional on-chain analysis tool we used and the available explorer views led us into the Hyperliquid infrastructure. However, they did not adequately answer the next crucial question:

What Happened to the Assets Next?

At precisely this point, the investigation had to move beyond conventional graphical on-chain analysis. Not because the trail had necessarily ended. But because we needed a different layer of data.

A missing line in a transaction graph does not automatically mean that the trail ends. Sometimes only the visibility provided by the tool has reached its limit.

From One Wallet to Twelve Relevant Account Structures

The further investigation quickly revealed why examining only the original destination account would not suffice. During the analysis, we examined a total of twelve Hyperliquid accounts or platform-specific account structures associated with the relevant asset movements.

These were not “off-chain accounts” in the conventional sense of a centralized exchange. However, the relevant events had to be examined at a Hyperliquid-specific data layer that was not adequately captured by the usual wallet-to-wallet view of our on-chain analysis tool. This also changed the investigative approach. The question was no longer simply:

Which Wallet Received the Stolen Crypto Assets Next?

Instead:

What Relationships Existed Between the Relevant Hyperliquid Accounts, What Internal Transfers Took Place, and Can the Economic Value Be Reconstructed Across These Movements?

This was the actual transition from conventional transaction tracing to platform-specific event reconstruction.

Why the Amounts Became Crucial

When reconstructing such movements, the size of individual amounts matters—but it must never be considered in isolation.

The incoming transfer under investigation comprised 70,500615 XMR1, with a USDC value of 35.151,606639 USDC reported by Hyperliquid at that time. Further comparison of the relevant account activities revealed a discrepancy of just 0,01 USDC in the target value—approximately one cent.

This discrepancy is significant from a forensic perspective.

Not because a difference of one cent mathematically proves that two events are identical. Rather, because in a multi-stage asset transformation followed by distribution across different account structures, such a small difference in value can constitute a very strong correlation indicator.

Put differently, for an examined volume of approximately 35.151,61 USDC, the difference was just one cent. Analytically, that is fundamentally different from a mere similarity between two roughly comparable amounts.

Nevertheless, the forensic limitation remains clear: amount correlation is an indicator—not conclusive proof on its own.

It must align with other characteristics, such as timestamps, account relationships, event types, and the sequence of movements.

When the Investigative Data Layer Changes

Professional blockchain analysis platforms are an essential component of modern cryptocurrency investigations. They enable, among other things, visualization of complex fund flows, identification of known services, exposure analysis, address attribution, and structured analysis of large volumes of blockchain data. are an essential component of modern cryptocurrency investigations. They enable, among other things, visualization of complex fund flows, identification of known services, exposure analysis, address attribution, and structured analysis of large volumes of blockchain data.

However, they cannot necessarily display every platform-specific activity at the same level of detail. As soon as relevant events are represented through dedicated account structures, ledger events, or specific transaction types, the investigative method must change as well. For us, this meant switching methods. Until then, the question had been:

Which On-Chain Transaction Follows the Known Asset Flow?

Now the question was:

Which Events Documented Within Hyperliquid Explain the Subsequent Movement of Economic Value?

Examining the Platform Data Directly

First, we analyzed different areas of Hyperliquid data separately. These included, in particular:

  • Non-Funding-Ledger-Updates
  • Spot Transfers
  • historical fill data
  • historical orders
  • account roles
  • and historical account and transaction information

The aim was not to collect as much data as possible. We wanted to distinguish between specific technical scenarios.

  • Had XMR1 merely been transferred between accounts?
  • Was there verifiable trading activity?
  • Were orders present?
  • Were there indications of another asset conversion?
  • Or did relevant movements lead to additional accounts within the same infrastructure?

What the Ledger Data Revealed

The first detailed ledger analysis of the central account under investigation returned 148 entries. Of these, 147 entries concerned XMR1 and one entry concerned USDC. The recorded events were classified as spot transfers. The known relevant incoming transfer of 70,500615 XMR1 was also documented as a spot transfer. The transaction data additionally confirmed a spotSend for exactly this amount.

This allowed us to verify the previously known asset flow at another technical data layer. At the same time, we had to account for the fact that this account did not exclusively move assets relevant to the case. The daily analysis showed substantial additional XMR1 activity.

In one dataset examined, for example, 110 outgoing XMR1 transfers to 102 different destination addresses were identified after the case-relevant incoming transfer. The total of these subsequent outgoing transfers exceeded 7.213 XMR1.

This made one thing clear: A simple approach based on the pattern 70,500615 XMR1 in → 70,500615 XMR1 out could not work here. The assets relevant to the case had entered a highly active account structure.

Why We Examined Trades Separately

Transfers and trades are not the same thing analytically. That is why we examined the corresponding data categories separately. For the relevant investigative context, the queried fill data produced no matching results. The historical orders likewise revealed no corresponding trading activity.

The role of the central account under investigation was also queried separately and identified as a regular user account. We were therefore able to establish the following: There was extensive transfer activity. However, the fill and order datasets we examined did not provide a direct explanation of how the relevant XMR1 might have been traded or converted during that period. This is an important distinction.

Why a Null Result Also Matters

An empty dataset does not automatically constitute a failed result in a forensic investigation. If transfer activity is visible while separately queried trading data shows no corresponding event, that discrepancy is analytically relevant as well. However, it does not justify replacing the missing explanation with speculation.

At this point, one could have constructed a plausible story. From a forensic standpoint, that would have been wrong. A technically plausible explanation is not proof. Events that have not been demonstrated must not be replaced with assumptions. That is why the investigation continued.

A Known Transfer as a Technical Anchor

For the subsequent reconstruction, we used the already verified transfer of 70,500615 XMR1 as a technical reference point. The event could be assigned to a specific historical block. This gave us a clearly defined starting point for the next stage of the investigation.

However, another problem then emerged. The standard account history did not contain the full historical depth we needed for the relevant period. The user details we queried covered only a limited selection of more recent transactions. We therefore could not simply scroll backward through the interface to the relevant point in time. We had to reconstruct the historical data specifically.

Why We Developed Our Own PowerShell Tool

Manually examining large historical block ranges would have been inefficient and prone to error. We therefore developed our own PowerShell-based analysis tool.

The objective was deliberately narrowly defined. The script was not intended to automate attribution of perpetrators or replace an existing blockchain analysis tool.

It was designed to answer one specific question: Which technically relevant events can be identified in Hyperliquid's historical data starting from the known reference point? For this purpose, we examined a defined historical block range.

The script

  • systematically retrieved historical block data
  • examined datasets for relevant account references
  • identified relevant transaction and event structures
  • logged matches
  • saved results in a structured format
  • accounted for rate limits
  • used controlled retries in the event of temporary errors
  • and saved its progress so that a scan could be resumed reproducibly

The need for this logic became apparent during operation: the interface returned HTTP status 429 for individual requests, prompting the script to pause and then continue in a controlled manner. This transformed a manual search into a reproducible forensic analysis process.

Why the Twelve Accounts Had to Be Examined Individually

Examining a platform structure involving multiple accounts differs fundamentally from tracing a linear wallet chain.

In a conventional blockchain transaction, it is often possible to establish directly:

  • Address A sends amount X to address B.

Within more complex infrastructure, however, several account relationships may become relevant in sequence. That is why we did not simply treat the twelve identified account structures as a single cluster.

For each account, we had to examine its role during the relevant period and determine which technically documented events existed.

The following were particularly important:

  • incoming and outgoing transfers
  • chronological sequence
  • amounts
  • USDC value representations
  • event types
  • account relationships
  • possible trading activity
  • and subsequent external movements

Only by combining these elements can individual technical events be assembled into a forensically defensible reconstruction.

Why One Cent Can Matter

The identified discrepancy of just 0,01 USDC demonstrates why forensic analysis cannot be reduced to searching for identical amounts. Conversions may involve rounding, price movements, fees, or different valuation times.

It would therefore be unrealistic to expect an asset to retain exactly the same numerical representation after several transformations. Conversely, a difference of just one cent in a five-figure amount is highly relevant analytically.

It can significantly strengthen a hypothesis if, at the same time,

  • the timing aligns
  • the account relationships can be traced
  • the asset sequence is technically plausible
  • and the corresponding platform events are present
The correct formulation is therefore not: “Because only one cent is missing, these are definitely the same assets.” Instead: “The discrepancy of just one cent, in conjunction with the other technical characteristics, provides strong evidence of correlation.” This distinction is essential to a forensically defensible conclusion.

From Transaction Tracing to Event Reconstruction

The case illustrates a fundamental difference between conventional blockchain analysis and investigations within complex trading infrastructure. A simple on-chain path might look like this:

  • Wallet A → Wallet B → Wallet C

The present case developed in a considerably more complex way:

  • Theft of ETH
  • ↓
  • Conversion to USDC
  • ↓
  • Conversion to XMR1
  • ↓
  • Transfer of 70,500615 XMR1
  • ↓
  • Hyperliquid
  • ↓
  • Analysis of twelve relevant account structures
  • ↓
  • Comparison of ledger, transaction, and historical data
  • ↓
  • further forensic reconstruction

Not every relevant intermediate step appears as a simple wallet-to-wallet line in a conventional on-chain analysis tool. At some point, therefore, merely following a transaction graph is no longer sufficient. Pure transaction tracing becomes event reconstruction.

Why We Had to Move Beyond the On-Chain Analysis Tool

Having to move beyond conventional on-chain analysis does not mean that the professional analysis tool we used was inadequate. Quite the opposite.

Without this analysis, we would not have been able to reconstruct the preceding asset path

  • ETH → USDC → XMR1 → Hyperliquid

in this form.

The problem had changed. We had moved from a conventional blockchain question to a platform-specific data question. The existing tools had led us to this technical infrastructure. For the next step, we needed additional layers of data.

Our PowerShell tool therefore continued the investigation where graphical on-chain analysis no longer provided sufficient historical granularity for this specific question.

Reproducibility Instead of Screen Observation

Another advantage of automated analysis lies in evidence preservation. For a robust forensic investigation, it is not enough for an analyst merely to have seen a dataset on a screen.

The underlying information must be preserved in as traceable a form as possible. We therefore stored relevant datasets in a structured format.

They can subsequently be

  • reviewed again
  • compared with new findings
  • reanalyzed technically
  • and, where appropriate, incorporated into forensic reports

The crucial question thus becomes not only:

What Did We Find?

But also:

How Did We Reach This Finding, and Can the Analytical Step Be Reproduced?

What Our In-House Tool Cannot Do

The limitations must also be clearly stated for custom-developed analysis tools.

Our script

  • does not automatically identify the beneficial owner of an account
  • does not establish a connection based solely on similar amounts
  • does not bypass technical access restrictions
  • does not access non-public customer data
  • and does not replace information that may only be obtainable from the platform or competent law enforcement authorities

The tool structures and examines technically accessible data. The forensic assessment remains the responsibility of the analyst.

No Single Tool Solves Every Case

This is one of the key findings of the investigation. Modern blockchain forensics is not simply a matter of opening one analysis tool and following a line until a case is solved.

In this case, the trail crossed several technical layers:

  • ETH
  • ↓
  • USDC
  • ↓
  • XMR1
  • ↓
  • professional on-chain analysis
  • ↓
  • Hyperliquid
  • ↓
  • platform-internal account analysis
  • ↓
  • historical data
  • ↓
  • PowerShell-based reconstruction

Each layer answered different questions. The decisive factor is therefore not using as many tools as possible. What matters is recognizing when a data source reaches its analytical limits and determining which layer must be examined next.

Closing Remarks

The investigation began with stolen ETH. The perpetrators first converted the assets into USDC and then into XMR1. A significant part of the trail led through a transfer of 70,500615 XMR1, with a USDC value of 35.151,606639 at the time, into the Hyperliquid infrastructure.

At that point, both the conventional explorer view and our professional on-chain analysis tool reached the limits of their analytical visibility for this specific question. Further investigation required analyzing twelve relevant Hyperliquid account structures and directly examining historical platform data.

Of particular relevance was a value correlation in which the discrepancy from the target amount was just 0,01 USDC. One cent is not proof. But for an examined value of approximately 35.000 USDC, such a discrepancy is far from meaningless when considered alongside timing, account relationships, and technical platform events.

That is the difference between a superficial transaction search and forensic reconstruction. We do not follow the most likely story. We follow the data.

And sometimes that means changing not just the tool but the very layer at which an asset must be examined.

Über den Autor

Albert Quehenberger is the founder and managing director of AQ Forensics GmbH and a specialist in blockchain forensics, cryptocurrency investigations, and digital asset analysis. With more than a decade of experience in intelligence, digital investigations, and blockchain analysis, he assists companies, banks, lawyers, authorities, and private individuals in investigating crypto fraud, tracing assets, and analyzing complex financial flows on public blockchains.

His professional background includes more than six years of service in military intelligence, a master's degree from the WU Executive Academy, specialist training for court-appointed experts, and training as an election observer in accordance with OSCE/ODIHR standards. He is also a member of the expert council of the European Institute for Future Generations (EIFG), a ChangeNOW Ambassador, and an adviser to Web3 education programs in Ethiopia.

<a href="https://www.linkedin.com/in/albert-quehenberger-6791a3174/" target="_blank"> Albert Quehenberger</a>
Founder & CEO AQ Forensics

Technical Documentation:

Contact / describe your case