Skip to main content
DFR Publication

Topic hub · risk

DeFi incident and liquidation log

This log names DeFi liquidations and exploits we have already covered — including Morpho’s 25 August 2026 PT-reUSD cascade (~$36 million of debt repaid, zero bad debt) and Resupply’s June 2025 wstUSR exploit (~$9.6 million). It is a skip-this checklist, not a live loss table and not a prediction. Isolation contains damage to a market; it does not save a looped borrower with a sub-3% health-factor buffer. Educational research, not financial advice.

Written by the Decentralized Finance editorial team, reviewed by Kaiser Khan. Not a live dashboard. Figures stay on the linked briefings.

Named incidents

liquidation Morpho / Pendle / Steakhouse

Morpho PT-reUSD cascade

A ~3% PT-reUSD oracle print (15-minute TWAP versus accretion curve) liquidated about $36.14 million of looped debt in 14 minutes. Lenders were reported whole. Not a hack.

Skip-this: Skip PT collateral and looping if you cannot name the oracle, LLTV, curator and Pendle-pool depth versus the Morpho book. Health factor ~1.03 is a 3% move from liquidation.

PT-reUSD Morpho briefing →

exploit Resupply

Resupply wstUSR donation / oracle

About $9.6 million taken from a wstUSR market live ~90 minutes. Donation inflated the share rate so collateral priced near zero. Insurance Pool took designed first-loss.

Skip-this: New, thin oracle markets are a different risk from ‘the protocol has been live a year’. Skip Insurance Pool yield if you have not read the post-mortem.

Resupply guide (incident section) →

exploit Kelp DAO

KelpDAO bridge incident

Our existing news coverage of the April 2026 KelpDAO bridge failure and restaking-token stress. Composability spread the shock into LST and lending books.

Skip-this: Bridges concentrate assets behind a smaller verification set than the base lending market. Isolation on Aave did not prevent withdrawals; it did prevent protocol-wide bad debt in that episode as we reported it.

KelpDAO incident analysis →

oracle Inverse Finance

Inverse Finance Anchor oracle manipulation

Two oracle-manipulation exploits on Inverse’s original Anchor market, combined losses over $24 million in contemporaneous coverage. Thin on-chain feeds, not a 2026 event.

Skip-this: Fixed-rate FiRM is a different product. Do not treat 2022 Anchor as a reason to skip 2026 FiRM without reading the current Inverse guide — and do not treat FiRM as Anchor either.

Inverse Finance guide →

Can I be looped out? — health-factor checklist

There is no live calculator here. A widget that pretended to know your Morpho health factor would lie the moment the oracle moved. Work these eight questions before you borrow against a yield token. If any answer is “I don’t know”, skip the position. Worked example: Steakhouse PT-reUSD/USDC on Morpho, 25 August 2026.

  1. 01

    What prices the collateral?

    Spot Chainlink, a 15-minute TWAP, an accretion curve, or ‘lower of two’. PT-reUSD used lower-of TWAP and a curve to $1 at maturity. If you cannot name the feed, skip the market.

  2. 02

    How large is the book versus the pool that prices it?

    A Morpho market several times the Pendle pool is a liquidation amplifier. Isolation stops contagion to other Morpho markets. It does not add liquidity to the pool you looped.

  3. 03

    What is LLTV / liquidation LTV?

    91.5% LLTV with a health factor of 1.03 is a sub-3% buffer. Aave eMode is the same family of trade: higher LTV, faster wipeout if correlation breaks.

  4. 04

    Are you looping?

    Deposit → borrow → buy more collateral → repeat. Advertised APY rises; the buffer shrinks. If a 3% wick liquidates you, you were not ‘earning yield’. You were selling a put to liquidators.

  5. 05

    Who is the curator?

    On Morpho, Steakhouse (or any named curator) sets which markets a vault uses. Curator pause/withdraw is a feature, not a backstop for your loop. If you cannot name them, skip the vault.

  6. 06

    Is this a new market?

    Resupply’s June 2025 wstUSR hole opened ~90 minutes after listing. Time-in-production on the parent protocol does not bless a brand-new oracle path.

  7. 07

    Can you add collateral in the same 14 minutes?

    The PT-reUSD cascade ran 04:37–04:51 UTC. Bots are faster than you. If your plan is ‘I’ll wake up and repay’, do not borrow.

  8. 08

    Have you supplied without borrowing first?

    If you have never completed a supply-only Aave deposit, skip Morpho PT markets, skip eMode, skip looping. The incidents log is not a beginner on-ramp.

Continue

FAQ

Incident log — FAQ

Was the August 2026 PT-reUSD event a Morpho hack?

No. Liquidations executed as the oracle and LLTV specified. Pendle and Steakhouse said the feed worked as designed. Borrowers with looped PT-reUSD lost collateral; vault lenders were reported whole with zero realised bad debt.

Does isolation make Morpho safer than Aave?

Safer for everyone else; not automatically safer for you. Isolation contains a bad market. Aave’s shared pool has different contagion and governance-listing trade-offs. Pick the venue for the job — see Morpho PT collateral vs Aave isolation.

Is this a live loss leaderboard?

No. Figures are from dated news and protocol post-mortems we already published. We do not invent a 2026 hack total. For historical bridge and lending exploits see Biggest DeFi hacks through 2025–26.

Where is the health-factor calculator?

There isn’t one on this site. A fake live widget would go stale. The checklist on this page is the product: oracle, LLTV, buffer, looping, curator, new-market risk.