Had some time to collect my thoughts…
Response to Recovery Plan: Full Position Restoration Required
Who I Am
I’m one of the 181 borrowers affected by the cbETH oracle incident from February 14–18. My wallet (0x3e2178...ad703b) had approximately $90,400 in collateral seized across 8 liquidation events, with only ~$42,880 in debt repaid — leaving me with a net loss of roughly $47,500. That was my life savings.
As of January 31, my supplied collateral was: 21.05 weETH, 18.93 cbETH, and 1.1 ETH — plus any interest accrued between then and the incident. At ~$2,200/ETH, that’s roughly $90,376 in collateral, which matches the $90,403.89 the remediation dashboard shows was seized. All of it — gone in 8 liquidation events over a matter of minutes, because of a bug in a governance proposal.
I maintained a healthy position. I did nothing wrong. This wasn’t market volatility. This wasn’t a bad bet. This was a protocol-level failure caused by a governance proposal that Moonwell itself pushed through.
In fact, just days before this incident, when ETH was dumping, I was actively protecting my position. I pulled thousands of dollars out of my bank account and sold other positions at a loss to cover my loan and keep my health factor safe. Real money, out of my own pocket, to be a responsible borrower on this protocol. That’s money I’ll never get back either — and it doesn’t show up in any remediation spreadsheet. I did everything a borrower is supposed to do, and then the protocol itself destroyed my position anyway.
Here is a screenshot of my positions from before I sold my AERO position to help cover the loan during the dump from Jan 31.
What Actually Happened
Let’s be very clear about the chain of events, because the framing matters:
-
MIP-X43 was a Moonwell governance proposal. It was not an external attack. It was not an oracle failure outside the protocol’s control. It was code that the DAO reviewed, approved, and executed.
-
The bug was elementary. The oracle used the raw cbETH/ETH exchange rate (~1.12) instead of multiplying it by ETH/USD to get the actual dollar price (~$2,200). SlowMist’s founder called it a “very low-level bug.” This should have been caught by any competent audit or review process.
-
The code was co-authored by an AI model without adequate human review. GitHub commit records show the oracle configuration logic was co-written by an AI. Using AI-generated code on the single most critical component of a lending protocol — the price oracle — without rigorous human verification is reckless.
-
There was no circuit breaker. Once the bad oracle went live, liquidations ran unchecked for days because fixing it required a 5-day governance vote and timelock. A protocol managing $90M+ in TVL had no emergency mechanism to pause liquidations from an obviously broken oracle.
-
This is the third oracle incident in six months. October 2025: AERO/VIRTUAL/MORPHO pricing discrepancy, $12M in liquidations, $1.7M bad debt. November 2025: wrsETH oracle malfunction, $3.7M bad debt. February 2026: cbETH, $2.68M in losses across 181 borrowers. This is not bad luck. This is a systemic failure in the protocol’s risk management and deployment practices.
Moonwell’s own post-mortem acknowledges that users “trusted Moonwell with their assets and were harmed through no fault of their own.” That statement is correct. Now the response needs to match the severity of the failure.
Why the Current Proposal Falls Short
1. $310K immediate against $2.68M in losses is not serious
That’s 11.5% upfront. For my position, that translates to roughly $5,500 now with the remaining ~$42K dependent on “future protocol revenue” over 12 months. You’re asking the people you harmed to bet on your future earnings. We already bet on your protocol once and lost everything.
2. Compensation in USDC forces a loss on top of a loss
If ETH appreciates over the next 12 months — which many of us were positioned for by holding ETH-denominated collateral — receiving USDC at February 2026 prices means we’re effectively being forced to sell our ETH at the worst possible time. We didn’t choose to exit our positions. The protocol liquidated us by mistake.
Compensation must be denominated in the assets that were seized — cbETH, not USDC. If the protocol cannot source cbETH, then WETH at minimum. Dollar-denominated repayment on an ETH-denominated loss is not making anyone whole.
3. Net loss calculation ignores the full picture
The methodology calculates: collateral seized (USD) minus debt repaid (USD). But many of us, myself included, had carefully structured positions with specific collateral compositions and borrowing strategies. I held 21.05 weETH, 18.93 cbETH, and 1.1 ETH — a deliberate mix of liquid staking derivatives chosen for their yield characteristics. We didn’t just lose a net dollar amount — we lost our entire positions, our collateral, and we’re still carrying bad debt.
True remediation means restoring positions: returning collateral AND restoring the corresponding debt positions so borrowers are back where they were on February 14. Anything less is a partial settlement dressed up as a recovery plan.
4. A 12-month Sablier vesting with expiry is unacceptable
The protocol makes a catastrophic error, wipes out users’ life savings, and then puts a deadline on when those users can claim their money back? If someone misses the window because they’re dealing with the fallout of losing their savings, they forfeit their claim? This is not how you rebuild trust.
What I’m Asking For
-
Full position restoration — my 21.05 weETH, 18.93 cbETH, and 1.1 ETH returned, with corresponding debt positions reconstructed to their pre-incident state as of February 14, 2026.
-
Compensation in-kind — cbETH or WETH, not USDC. We held ETH-denominated assets for a reason.
-
Accelerated timeline — the protocol has access to the Aera funds and other treasury resources. Waiting 12 months for revenue to trickle in while users sit with wiped accounts is not accountability.
-
No expiry on claims — affected users should be able to claim remediation without a ticking clock.
-
Structural changes — emergency pause mechanisms, mandatory independent audits for all oracle changes, and an end to deploying AI-generated code on critical infrastructure without multi-party human review.
Final Thoughts
I’ve used Moonwell as a core part of my DeFi strategy. I trusted this protocol. I maintained a healthy position. I did everything right — including spending thousands of dollars out of my own bank account days before the incident to keep my loan safe during a market dip. And I lost my life savings because of a bug that should never have made it to production.
The net loss calculation doesn’t capture the full damage. It doesn’t account for the money I pulled from my bank to shore up my position. It doesn’t account for the other positions I sold at a loss to cover my loan. Those are real, additional losses that happened specifically because I was being a responsible borrower on your platform — losses that are now compounded by the fact that the position I was protecting got wiped out by the protocol itself days later.
The community is watching how Moonwell handles this. A protocol that makes its users whole after its own mistake earns long-term trust and loyalty. A protocol that offers partial reimbursement in the wrong denomination on a 12-month maybe-timeline with an expiry date does not.
181 borrowers. $2.68 million. This is not an abstract governance exercise — these are real people who lost real money because of a real failure in your process. Act accordingly.
bad debt is one thing but user funds should be covered immediately!
