An attacker exploited Resolv's USR minting mechanism to create 80M unbacked tokens with ~$200K in capital, extracting ~$25M. A second, distinct attack vector emerged: Morpho's supplyOnBehalf function allowed the attacker to force-allocate USDC into compromised markets, bypassing vault-level supply caps entirely — a protocol-level design issue disclosed by Re7 Labs. The outcome now hinges on two paths: if RLP absorbs the loss, USR repegs to $1 but RLP holders take a 65% haircut and Morpho retains ~$5M in structural wstUSR bad debt; if RLP does not absorb, USR stays at $0.33–$0.75 and total lending market bad debt reaches $10–20M. Critically, Fluid's 6x leveraged wstUSR positions (98% of wstUSR supply) remain underwater in both scenarios because the wstUSR wrapping ratio (0.1725) creates structural insolvency that a USR repeg cannot fix. Fluid's exposure likely exceeds all Morpho bad debt combined.
| Time (UTC) | Event | Impact |
|---|---|---|
| ~02:21 | First mint — 50M USR via requestSwap ($100K USDC deposit) | 500:1 return; unbacked supply injected |
| ~02:25 | Second mint — 30M USR | Total unbacked supply reaches 80M tokens |
| ~02:30 | Attacker begins dumping USR across DEXs | Massive sell pressure on Curve, Uniswap pools |
| ~02:38 | USR hits $0.025 on Curve | 97.5% depeg; Curve pool drained of stablecoins |
| ~02:46 | Re7 Labs internal monitoring flags exploit | 25 minutes after first mint; automated alert triggers |
| ~03:00 | Re7 sets caps to 0 on Base vault, removes impacted markets | Supply queues cleared on both affected Morpho markets |
| ~03:02 | Re7 sets caps to 0 on mainnet vault | All Re7-curated USR exposure neutralized |
| Hours after | Resolv pauses all protocol functions | Minting, redemptions, and swaps halted |
| Post-pause | Re7 discloses Morpho supplyOnBehalf donation attack to Morpho team |
Second attack vector identified at protocol level |
| Post-pause | Morpho advises all curators to remove markets from supply queues | Mitigation for supplyOnBehalf bypass |
| Ongoing | USR at $0.33; wstUSR at $0.057; RLP at $0.82 | All USR/wstUSR Morpho markets frozen at 100% utilization |
The SERVICE_ROLE was a single externally owned account, not a multisig. The attacker gained control of this key. No oracle validation, no amount caps, no rate limiting existed on the minting path. The attacker deposited ~$100K USDC via requestSwap and received 49.95M USR — a 500:1 return. A second transaction minted an additional 30M USR. The attacker then dumped USR across DEX pools, converting proceeds to 11,409 ETH (~$23.7M). As on-chain analyst Vadim (@zacodil) noted: this was not a smart contract bug but a design flaw. The requestSwap function was designed to trust the SERVICE_ROLE completely.
A separate donation attack exploited Morpho's supplyOnBehalf mechanism. The attacker force-allocated USDC into the USR/USDC market by calling Morpho's underlying contract directly on behalf of a vault address. This bypassed the vault's supply queue and supply cap controls entirely — restrictions that operate at the vault layer do not prevent direct calls to the Morpho contract. Re7 Labs disclosed this to Morpho, who advised all curators to remove markets from supply queues as a mitigation. This vulnerability is distinct from the USR mint exploit and represents a Morpho protocol-level design issue.
Permissionless composability becomes an attack surface under stress. The supplyOnBehalf function was designed for legitimate use — third parties depositing on behalf of vaults. But during a depeg, an attacker can force-deposit into a compromised market, inflating the vault's exposure beyond what the curator intended. Since supply caps only govern vault-level allocation decisions (not underlying Morpho contract calls), the attacker circumvented all risk controls. This effectively turned Morpho's permissionless composability into an attack surface.
Re7 Labs' flagship mRe7YIELD vault has zero exposure. Of 184 curated markets, only 2 small Morpho markets where USR is listed as collateral were affected. Re7's internal monitoring flagged the exploit at 2:46am UTC — 25 minutes after the first mint. By 3:00am UTC, Re7 set all caps to zero and removed impacted markets from supply queues on both Base and mainnet.
USR at $0.33, wstUSR at $0.057, RLP at $0.82. USR's 66.7% discount to peg reflects the market's uncertainty about whether RLP will absorb the loss. wstUSR's 94.3% collapse is amplified by the wrapping ratio (0.1725 USR per wstUSR), meaning wstUSR is priced at roughly $0.33 x 0.1725 = $0.057 — consistent with a fully impaired USR. RLP's 18% discount suggests the market believes partial-to-full loss absorption is likely, but the outcome remains binary.
Six markets carry confirmed bad debt, all frozen at 100% utilization. The table below reflects actual on-chain state. Bad debt was created during the depeg window when arbitrageurs posted discounted USR/wstUSR at inflated oracle valuations and extracted USDC. These positions are now locked.
| Market | Chain | Collateral | LLTV | Supply ($) | Borrow ($) | Collateral ($) | Bad Debt ($) |
|---|---|---|---|---|---|---|---|
| 0xd9e3... | ETH | wstUSR | 91.5% | 6,225,488 | 6,225,488 | 1,273,911 | 4,951,577 |
| 0xff0f... | Base | USR | 91.5% | 1,243,210 | 1,243,210 | 549,871 | 693,339 |
| 0x8e7c... | ETH | USR | 91.5% | 251,726 | 251,726 | 97,676 | 154,050 |
| 0xe1b6... | ETH | RLP | 86.0% | 2,416,026 | 2,416,026 | 13,304,348 | 0 |
| 0x1cfd... | ETH | PT-RLP | 86.0% | 482,427 | 482,427 | 416,111 | 66,316 |
| 0x4b86... | ETH | RLP | 86.0% | 315,922 | 315,922 | 255,211 | 60,711 |
| Others | Multi | Various | — | 56,824 | 56,824 | 50,333 | 19,177 |
| TOTAL | $11,017,623 | $11,017,621 | $5,945,170 | ||||
98% of wstUSR supply locked in Fluid at 6x leverage. Per Euler governance forums, nearly all wstUSR was deployed into Fluid leverage loops. With wstUSR at $0.057, these positions are catastrophically underwater. The $5.2M in 24-hour wstUSR/USDC volume on Fluid suggests active position unwinding. Fluid's total bad debt likely exceeds all confirmed Morpho losses combined — estimated $15–25M depending on total wstUSR locked.
Venus Flux is a Fluid fork, and its USR markets are suspended. Venus Core on BNB Chain is unaffected because USR was not accepted as collateral in the main deployment. The exposure is isolated to the Ethereum-side Flux markets. Venus was already managing $2.15M in bad debt from a separate THE token flash loan attack on March 15 — one week before the USR depeg. The compounding effect of two simultaneous bad debt events strains Venus's risk fund.
USDC depositors in affected vaults cannot withdraw. Every Morpho market accepting USR or wstUSR as collateral is at 100% utilization. Recovery depends on borrowers voluntarily repaying or liquidations occurring, neither of which is likely while USR remains depressed. The only healthy market is RLP/USDC (0xe1b6), where RLP's $13.3M collateral substantially exceeds the $2.4M in borrows.
The total loss is fixed at $25M — already sitting in the attacker's wallet as ETH. The analytical question is not "what if USR trades at price X" but "how does the $25M loss flow through Resolv's capital structure to reach end depositors." Everything downstream — Morpho bad debt, Fluid insolvency, depositor losses — depends on which path Resolv takes.
Resolv burns 80M unbacked USR and charges $25M to RLP. The RLP pool drops from $38.6M to $13.6M — a 65% haircut. USR returns to $1.00 peg. RLP holders bear the entire loss. This is what RLP was designed for: it is the junior tranche / first-loss capital in Resolv's structure, backed by excess collateral in the treasury (ETH/BTC delta-neutral positions). Resolv's claim that the "collateral pool is intact" supports this path.
| Pre-exploit RLP pool | $38.6M |
| Loss absorbed by RLP | $25.0M |
| Post-loss RLP pool | $13.6M |
| RLP haircut | ~65% |
| Implied USR value | $1.00 (full repeg) |
wstUSR/USDC (0xd9e3): $4.95M bad debt persists — structural, not fixable. Even at USR = $1.00, wstUSR reprices to ~$0.1725 (the wrapping ratio). The $1.27M of wstUSR collateral cannot cover $6.2M borrowed. This bad debt was created during the depeg window when borrowers extracted USDC against wstUSR posted at $1 oracle valuation. The wrapping ratio means the collateral can never cover the borrows. USDC depositors in this market lose ~79.5% regardless of repeg.
USR/USDC markets (0xff0f, 0x8e7c): Bad debt approaches $0 if unwound properly. If USR repegs to $1, collateral is worth face value again. Borrowers' positions become collateralized. USDC depositors can potentially recover if borrowers are liquidated at par. However, the borrowers profited from the depeg-window arb — they extracted USDC at $1 against USR bought at $0.40. They have no economic incentive to repay. Recovery depends on liquidation mechanics, not voluntary repayment.
RLP/USDC (0xe1b6): No bad debt. RLP at $0.82 with a 65% haircut implies a post-loss RLP price of ~$0.35 x pre-exploit value. But the $13.3M of RLP collateral against $2.4M borrowed means the market remains overcollateralized even with the haircut. At 86% LLTV, this position is safe.
Net Morpho bad debt in Path 1: ~$5M. The wstUSR/USDC structural deficit drives the total. USR-backed markets recover partially. RLP markets are unaffected.
wstUSR leverage positions remain underwater even with a full USR repeg. If USR repegs to $1.00, wstUSR = $1.00 x 0.1725 = $0.1725. At 6x leverage, borrowers borrowed ~5x their collateral value. A position that started at $1.00 wstUSR with 6x leverage has $5 of debt against collateral now worth $0.1725 — 97% underwater even after repeg. The wrapping ratio is the problem, not the USR price. Path 1 does not save Fluid's wstUSR positions.
Fluid bad debt estimate in Path 1: $15–20M. If $50M+ of wstUSR was locked in Fluid (98% of supply), even at 50% utilization with 6x leverage, the gap between collateral value ($0.1725 per wstUSR) and borrowed stablecoins is enormous. Samyak's "first loss goes to RLP" applies to Resolv's internal accounting — RLP absorbs the collateral pool loss. But Fluid's wstUSR leverage pools are a separate layer. The RLP absorption fixes USR, not wstUSR leverage positions.
Venus Flux mirrors Fluid's mechanics as a Fluid fork. Its USR market suspension suggests similar exposure patterns. If USR repegs, Venus Flux's direct USR positions recover. But any wstUSR leverage exposure follows the same structural insolvency as Fluid. Venus Core is unaffected.
| USR holders | Made whole (repeg to $1.00) |
| RLP holders | 65% loss ($25M haircut on $38.6M pool) |
| Morpho USDC depositors (wstUSR vaults) | ~79.5% loss (structural, irreducible) |
| Morpho USDC depositors (USR vaults) | Partial recovery possible via liquidations |
| Morpho USDC depositors (RLP vaults) | No loss (overcollateralized) |
| Fluid USDC depositors | Significant loss ($15–20M, wstUSR leverage) |
| Venus Flux depositors | Partial recovery on USR positions; wstUSR impaired |
| Arbitrage traders | Keep their depeg-window profits |
Resolv does not use RLP to absorb, or cannot. The $25M loss sits with USR holders as dilution. Orderly haircut: $25M / ~$100M circulating = 25%, implying USR = $0.75. Confidence collapse scenario: USR stays at $0.33 or lower as holders rush to exit. RLP maintains its value — the insurance layer does not pay out, a perverse outcome for a first-loss tranche.
| RLP absorption | $0 (protocol cannot or will not use RLP) |
| USR haircut (orderly) | $25M / $100M = 25% |
| Implied USR (orderly) | $0.75 |
| Implied USR (confidence collapse) | $0.33 or lower (current market) |
USR/USDC markets: Collateral at $0.75 x posted amount, positions undercollateralized. At 91.5% LLTV, the borrowable amount exceeds the impaired collateral value. Bad debt = borrowed - (collateral x 0.75 x LLTV). Estimated bad debt: ~$350K across both USR/USDC markets. Not catastrophic, but permanent.
wstUSR/USDC (0xd9e3): Even worse. wstUSR = $0.75 x 0.1725 = $0.129. Collateral worth ~$900K against $6.2M borrowed. Bad debt: ~$5.3M. The wrapping ratio amplifies the loss non-linearly.
RLP markets: RLP holds value, no bad debt. If RLP does not absorb, it maintains its pre-exploit NAV. The $13.3M of RLP collateral remains healthy. Paradoxically, the market where the insurance token sits is the safest.
Net Morpho bad debt at $0.75: ~$5.6M.
This is what Path 2 looks like right now. The $5.95M in confirmed on-chain bad debt reflects Path 2 pricing. If USR remains at $0.33, no recovery is possible in the USR/wstUSR markets. The entire $5.95M persists.
wstUSR = $0.33 x 0.1725 = $0.057 (current price). At 6x leverage on $0.057 wstUSR, positions are completely wiped. Collateral is worth pennies against dollar-denominated stablecoin debt. Fluid bad debt: potentially $15–25M depending on total wstUSR locked. The $5.2M 24-hour volume suggests active liquidation and panic selling, but much of the damage is already irreversible.
| USR holders | 25–67% loss (depending on confidence) |
| RLP holders | Preserved (insurance layer doesn't pay out) |
| Morpho USDC depositors (all vaults) | $5.95M+ bad debt persists in full |
| Fluid USDC depositors | $15–25M bad debt (wstUSR leverage wiped) |
| Venus Flux depositors | USR positions impaired, no recovery |
| Resolv team / reputation | Destroyed — first-loss tranche failed its purpose |
| Market | Oracle Type | Path 1 (USR $1.00) | Path 2 Orderly ($0.75) | Path 2 Collapse ($0.33) |
|---|---|---|---|---|
| wstUSR/USDC (0xd9e3) | Hardcoded $1 | $4.95M | $5.3M | $4.95M |
| USR/USDC Base (0xff0f) | Hardcoded $1 | ~$0 | ~$311K | ~$693K |
| USR/USDC ETH (0x8e7c) | Hardcoded $1 | ~$0 | ~$63K | ~$154K |
| RLP/USDC (0xe1b6) | Market price | $0 | $0 | $0 |
| PT-RLP/USDC (0x1cfd) | Redemption | ~$0 | ~$66K | ~$66K |
| RLP/AUSD (0x4b86) | Market price | ~$0 | ~$61K | ~$61K |
| TOTAL MORPHO | ~$5.0M | ~$5.8M | ~$5.9M | |
| FLUID (EST.) | $15–20M | $15–25M | $15–25M | |
| VENUS FLUX (EST.) | $1–3M | $2–5M | $3–7M | |
| TOTAL LENDING BAD DEBT | ~$21–28M | ~$23–36M | ~$24–38M | |
The critical insight: Fluid's wstUSR leverage makes total lending bad debt roughly equal in both paths. Whether USR repegs or not, the wstUSR wrapping ratio (0.1725) ensures that leveraged wstUSR positions in Fluid are underwater. The only difference between paths is who bears the Resolv-level loss (RLP holders vs. USR holders) and whether USR/USDC Morpho markets recover. The wstUSR damage is done.
98% of wstUSR supply locked at 6x leverage created a $50M+ exposure. Per Euler governance forums, nearly all wstUSR was deployed into Fluid lending pools where users could borrow up to 12.5x on Ethereum Mainnet. At 6x average leverage, for every $1 of wstUSR deposited, users borrowed ~$5 in stablecoins. When wstUSR collapsed 94% from $1.00 to $0.057, these positions became catastrophically underwater. The leverage amplified losses by 6x: a 94% price drop on 6x leverage means the collateral is worth 0.36% of the debt.
Morpho's total quantified bad debt is $5.95M across $11M of supply — Fluid's could be 3–4x larger. If $50M+ of wstUSR was in Fluid (98% of supply), even at 50% utilization, the stablecoin borrowed against that collateral could reach $25M+. With wstUSR now worth $0.057 per token, the collateral backing those borrows has evaporated. The 24-hour wstUSR/USDC volume of $5.2M on Fluid confirms massive position unwinding, but Fluid has no public API, making precise quantification impossible. The structural exposure through 6x leveraged wstUSR is conservatively estimated at $15–25M in bad debt.
The RLP first-loss mechanism operates at the Resolv protocol layer, not at the Fluid lending layer. Samyak Jain's statement that "first loss capital should go to RLP" is correct for Resolv's internal accounting: RLP absorbs losses in the collateral pool, which could allow USR to repeg. But Fluid's wstUSR leverage pools are a separate layer of the stack. Even if RLP absorbs the full $25M exploit loss and USR repegs to $1.00, the wstUSR wrapping ratio (0.1725) means each wstUSR token is worth only $0.1725 — not $1.00. Positions that borrowed $5 of stablecoins per $1 of wstUSR are still 97% underwater. The RLP absorption fixes USR, not wstUSR leverage.
Fluid acted within approximately 30 minutes of the first depeg signal. Between roughly 02:30 and 02:50 UTC, USR was trading at $0.15–$0.40 on DEXs but had not yet been paused on Fluid. Opportunistic borrowers could have posted depegged USR/wstUSR as collateral during this window. The admin team paused new borrowings, froze new deposits, and began analyzing existing positions. The volume of new borrowings during this window determines the incremental damage beyond pre-existing leverage positions.
Venus Flux is a Fluid fork with the same architectural exposure patterns. Its USR market suspension suggests the Venus team recognized the contagion risk immediately. The key difference: Venus Core markets on BNB Chain are unaffected, and Venus's risk fund was already strained by the $2.15M THE token exploit from March 15. Managing two simultaneous bad debt events — THE and USR — limits Venus's capacity to backstop Flux losses. Venus Flux's bad debt is estimated at $1–7M depending on the path Resolv takes.
RLP is the junior tranche that absorbs losses before USR holders are impaired. RLP represents excess collateral beyond $1-per-USR backing. If an epoch ends with a negative return, 100% of that loss is deducted from the RLP pool. Pre-exploit RLP pool: $38.6M, backed by ETH/BTC delta-neutral hedges (basis trades). The $25M extraction represents approximately 65% of the entire pool. RLP's current market price of $0.82 implies the market values the post-loss pool at ~$31.6M — suggesting partial but not full loss absorption is priced in.
Stream Finance holds 13.6M RLP on Morpho, approximately $17M net exposure. Stream already disclosed a $93M loss in November 2025 from the xUSD depeg/insolvency. Its depositors face yet another significant hit. Platforms like yoUSD that use RLP as collateral face knock-on effects. Stream's concentrated RLP position means a 65% RLP haircut would destroy approximately $11M of its Morpho collateral value.
Resolv raised $10M from Cyber.Fund, Maven11, Coinbase Ventures, and Arrington Capital. Community speculation centers on whether these VCs would inject additional capital to restore confidence. The prevailing view in the Resolv Discord: "$25M isn't hefty to restore confidence" — suggesting some expect a capital raise or VC backstop. However, no VC has made a public commitment. The RLP pool's $38.6M pre-exploit size means RLP alone has sufficient capacity to absorb the loss without external capital, if Resolv chooses to deploy it.
Resolv's profit distribution documentation states: "If the protocol realizes a loss over such 24-hour period, it is allocated to RLP." The wording is unambiguous — RLP bears 100% of downside. For exchange insolvency scenarios, the docs specify that losses are "deducted from the collateral pool's value" and "hit RLP exclusively," with "USR remains fully redeemable from remaining on-chain assets." This language strongly supports Path 1 — RLP should absorb the $25M loss.
The documented loss scenarios cover negative funding rates, exchange counterparty failures, and market volatility — all situations where the collateral pool loses value. This exploit is fundamentally different: the collateral pool is intact (per Resolv's own statement). The $25M wasn't lost from the pool — it was created from nothing via unbacked minting and extracted. The question is whether "protocol realizes a loss" extends to exploit-generated unbacked supply dilution. The docs are ambiguous on this distinction.
The Steakhouse Financial analysis of Resolv notes: "In the absolute worst case where RLP goes to 0 and funding is still negative, USR could become undercollateralized — essentially a failure of the stablecoin." With RLP at $38.6M pre-exploit and the extraction at ~$25M, RLP would survive at ~$13.6M (65% haircut) in Path 1. RLP is not wiped — but it enters a "thin protection layer" state where any further adverse event could push it to zero.
The protocol suspends RLP redemptions when the USR collateralization ratio drops below 110%. This mechanism is likely active now. If CR is below 110%, RLP holders cannot exit — they are locked in as the loss-absorbing layer. This is by design but creates a reflexive dynamic: locked-in RLP holders cannot flee, which stabilizes the system but at the cost of RLP holder liquidity.
Steakhouse's Monte Carlo simulation of 500 paths concluded "the chances of the entire RLP pool being wiped are virtually 0." But this analysis used historical 2025 funding and staking data — it did not model a scenario where $80M of unbacked tokens are minted via a key compromise. The tail risk that materialized was operational (private key security), not market-driven. No Monte Carlo simulation captures operational risk of this nature.
Because the docs don't explicitly address exploit-generated losses vs market losses, the Resolv team has discretion. If they choose to honor the RLP-first-loss structure (Path 1), USR holders are protected but RLP holders take a 65% haircut. If they argue the exploit falls outside RLP's coverage (Path 2), USR holders take the hit. The team's incentive is likely Path 1 — burning the unbacked USR and letting RLP absorb, because this preserves USR's peg and the protocol's viability. Path 2 kills the stablecoin.
The $10M seed round included Coinbase Ventures, Cyber.Fund, Maven11, and Arrington Capital. These investors have reputational stakes in the stablecoin narrative. Community speculation on Discord: "$25M isn't a hefty price to pay to keep the stablecoin concept alive." VC backstop funding could supplement RLP's absorption, reducing the haircut below 65% and accelerating confidence recovery.
The SERVICE_ROLE should have been a multisig or timelock. No protocol managing $100M+ in TVL should operate with a single-key privileged role that can mint unlimited tokens. A 3-of-5 multisig with hardware wallet signers would have made key compromise insufficient to execute the attack. This is the same failure mode as Ronin Bridge, Harmony Horizon, and Multichain.
A single require() statement would have prevented the exploit. The minting function should have validated that the USR amount being minted corresponded to the USDC deposit within a tight tolerance band (1:1 +/- 0.5%). An on-chain Chainlink or TWAP oracle check would have flagged the 500:1 mint ratio as anomalous and reverted automatically.
Hardcoded $1 oracles for USR/wstUSR created the entire arbitrage surface. Morpho vault curators valued USR at a constant $1 regardless of actual market price. The gap between $1 face value and $0.40 market price was a deliberate design choice by curators who prioritized simplicity over resilience. Market-price oracles with liquidity-weighted averaging would have reflected the depeg within minutes.
The 0.1725 wrapping ratio is a leverage multiplier that most risk models missed. When USR depegs, wstUSR doesn't fall 1:1 — it falls by the depeg percentage applied to an already-discounted token. At USR = $0.33, wstUSR = $0.33 x 0.1725 = $0.057, a 94.3% decline. Combined with 6x leverage in Fluid, the effective loss amplification is ~34x. This non-linear relationship between USR price and wstUSR collateral value was not adequately modeled by any lending market that accepted wstUSR.
Independent Research — March 22, 2026