KVORUM
aaveSource: Aave governance v3 · #482

rsETH Incident: WETH LTV Restoration Across V3 Instances

executedMay 15, 2026, 05:53 PM UTCby660x66a2…73ff

01Summary

AISummary by Kvorumclaude-haiku-4-5

Restore WETH Loan-to-Value (LTV) parameters across Aave V3 instances on Ethereum Core, Ethereum Prime, Arbitrum, Base, Mantle, and Linea to pre-exploit levels (ranging from 80-84% base LTV and 90-93% E-Mode LTV). This follows successful recovery of 106,993 rsETH from the April 2026 exploit and secured DeFi United coalition funding to cover remaining bad debt.

Type
parameter change

Key changes

  • Restore Ethereum Core WETH base LTV from 0% to 80.50%(high)
  • Restore Ethereum Prime WETH base LTV from 0% to 84.00%(high)
  • Restore Arbitrum WETH base LTV from 0% to 80.00%(high)
  • Restore Base WETH base LTV from 0% to 80.00%(high)
  • Restore Mantle WETH base LTV from 0% to 80.50%(high)
  • Restore Linea WETH base LTV from 0% to 80.00%(high)
  • Restore Ethereum Core WETH E-Mode LTV from 0% to 93.00%(high)
  • Restore Arbitrum WETH E-Mode LTV from 0% to 93.00%(high)

Notable concerns

  • Restoration is contingent on rsETH recovery efforts progressing as planned
  • Bad debt coverage from DeFi United coalition must be successfully deployed
  • Remaining ~5,211 rsETH sold on DEXs requires external coalition funding to cover
  • Arbitrum court-authorized $71M stolen ETH transfer must complete as scheduled
ContractsAaveV3EthereumAaveV3EthereumLidoAaveV3ArbitrumAaveV3BaseAaveV3MantleAaveV3Linea
Generated by Kvorum
model: claude-haiku-4-5
provider: anthropic
profile: anthropic-fast-v1
prompt: v1.0
generated: 2mo ago
Read the description ↗

02Description

662 words

Simple Summary

Restore WETH Loan-to-Value (LTV) on Aave V3 Ethereum Core, Ethereum Prime, Arbitrum, Base, Mantle, Linea to their respective pre-exploit values. The Risk Guardian has already lifted the WETH freeze on the L2 instances where it remains in place. Together, these actions return the WETH market to normal operation across all Aave V3 deployments affected by the precautionary measures applied during the 2026-04-18 Kelp/LayerZero OFT bridge exploit.

Motivation

The temporary WETH freeze on L2s and LTV=0 across all affected instances were precautionary measures applied at the time of the 2026-04-18 exploit, alongside the freeze of rsETH and wrsETH reserves. The recovery has since progressed substantially, as outlined in the Aave update of 2026-05-08:

  • On 2026-05-06, the attacker positions across Aave V3 Ethereum Core and Arbitrum were liquidated via a specifically constructed AIP payload.
  • Combined with Compound’s parallel liquidations, 106,993 rsETH has been recovered (89,567 from Aave V3 + 17,426 from Compound) out of an estimated 112,103 rsETH of unbacked supply on the affected L2s. The retrieved rsETH was transferred to the Aave Recovery Guardian.
  • The remaining ~5,211 rsETH (sold by the attacker on DEXs and not recovered through liquidations) is to be covered by the DeFi United coalition’s committed ETH.
  • Mantle DAO has formally joined DeFi United via onchain governance vote.
  • A New York court has authorised an onchain Arbitrum DAO vote to transfer the immobilised $71M of stolen ETH (recovered by Arbitrum’s Security Council) to Aave LLC for the recovery effort.

Given the secured DeFi United funds and the demonstrated recovery progress, the precautionary WETH constraints on Aave can be lifted now without compromising user protection. Holding WETH LTV at 0 longer than necessary creates avoidable friction for end-users. This advisory to unfreeze and restore WETH LTV is contingent on the rsETH restoration efforts progressing as planned and with the successful deployment of DeFi United funds to cover the bad debt in all affected Aave markets.

Specification

Base Reserve LTV Restoration

Restore the WETH base reserve LTV to its pre-2026-04-18 value on each Aave V3 deployment that received the precautionary LTV=0 as part of the reserve freeze enacted via the Aave Protocol Guardian. Liquidation Threshold (LT), Liquidation Bonus (LB), Reserve Factor, and the usageAsCollateralEnabled flag remain unchanged.

Aave V3 InstanceCurrent LTVTarget LTV (pre-exploit)LT (unchanged)LB (unchanged)
Ethereum Core0.00%80.50%83.00%5.00%
Ethereum Prime0.00%84.00%85.00%5.00%
Arbitrum0.00%80.00%84.00%5.00%
Base0.00%80.00%83.00%5.00%
Mantle0.00%80.50%83.00%5.50%
Linea0.00%80.00%83.00%6.00%

E-Mode LTV Restoration

The precaution also implicitly set the LTV-zero flag on WETH in the ETH-correlated E-Mode of each market where WETH is enabled as E-Mode collateral. The flag zeroes the E-Mode LTV for the flagged asset without modifying the eMode category configuration or the collateral bitmap; LT and LB inside the eMode are unaffected. Since Aave v3.6, this state can be reset via setAssetLtvzeroInEMode.

MarketE-Mode IDCategoryLTVLT (unchanged)
Ethereum Core1ETH correlated93.00%95.00%
Arbitrum2ETH correlated93.00%95.00%
Base1ETH correlated90.00%93.00%

After the flag is unset on these markets, the effective WETH E-Mode LTV reverts to the eMode category cap: 93% on Ethereum Core and Arbitrum, 90% on Base.

References

Copyright

Copyright and related rights waived via CC0.

03Mismatch analysis

AIMismatch analysis by Kvorumlow confidenceclaude-sonnet-5
Consistent

The decoded calldata consists of only two generic execute() (fallback) calls to payload contracts on Ethereum mainnet, with no decoded internal arguments. This is the standard pattern for Aave 'Direct-to-AIP' governance proposals, where the mainnet vote triggers execute() on one or more payload contracts whose source code (linked in the References section) contains the actual per-market LTV and E-Mode parameter changes, and which may internally trigger cross-chain governance actions to reach the L2 deployments (Arbitrum, Base, Mantle, Linea) mentioned in the description. Because the calldata here is opaque (no parameter-level decoding), I cannot independently verify that the specific LTV values (80.50%, 84.00%, 80.00%, etc.) or the E-Mode LTV-zero flag resets described in the prose are exactly what the two execute() calls perform. However, there is nothing in the decoded data that contradicts the description, and the number and pattern of calls (two mainnet payload executions, consistent with Core + Prime, with the other chains reached via cross-chain messaging inside those payloads) is consistent with how Aave typically structures these multi-chain risk parameter proposals. Given the lack of granular calldata to confirm exact values, confidence is low, but I found no basis to flag a discrepancy — the surface-level structure matches expectations for this type of proposal.

Generated by Kvorum
model: claude-sonnet-5
provider: anthropic
profile: anthropic-strong-v1
prompt: v1.0
generated: 2mo ago
Review the decoded actions ↗

04Actions

2 total
#0fallback()chain 0x1
Target
0x3dc4…4ca6
Value
0 wei
Function
fallback()
Arguments
{}
Calldata
#1fallback()chain 0x1
Target
0x38c5…5e3e
Value
0 wei
Function
fallback()
Arguments
{}
Calldata

05Tally

Summed from votes

Voters0
VP participating0
Outcome (current)—
Current support—

06Voters

0 total

No votes recorded yet.

07Forum

No forum thread is confidently linked to this proposal.