KVORUM
aaveSource: Aave governance v3 · #270

Clinic steward activation

executedMar 18, 2025, 10:00 AM UTCbyF70xf71f…1e02

01Summary

AISummary by Kvorumclaude-haiku-4-5

Aave DAO proposes enabling a new ClinicSteward smart contract to clean up approximately $400,000 in legacy bad debt accumulated since v3 activation. The contract will be controlled by a 2/3 multisig (ACI/TokenLogic/Karpatkey) to liquidate and repay unhealthy positions using Aave Collector funds, with pre-configured spending caps per network ranging from $1,000 to $350,000.

Type
contract upgrade
Funding
$$500000

Key changes

  • Deploy and activate ClinicSteward smart contract across multiple networks to facilitate bad debt cleanup(high)
  • Grant ClinicSteward the FUNDS_ADMIN role in Aave DAO(high)
  • Grant ACI-controlled EOA/Bot the CLEANUP_ROLE to execute liquidations and repayments(high)
  • Implement per-network spending caps for bad debt cleanup (total approximately $500,000 across all networks)(medium)

Notable concerns

  • Grants significant authority to ACI-controlled bot through CLEANUP_ROLE
  • Relies on 2/3 multisig (ACI/TokenLogic/Karpatkey) oversight of cleanup operations
  • Authorizes access to Aave Collector funds for liquidation activities
ContractsClinicStewardAave CollectorAave DAO governance
Generated by Kvorum
model: claude-haiku-4-5
provider: anthropic
profile: anthropic-fast-v1
prompt: v1.0
generated: 2mo ago
Read the description ↗

02Description

309 words

Simple Summary

Enabling a new Aave ClinicSteward smart contract, with capabilities to liquidate and repay non-healthy positions. The contract will utilize Aave Collector funds for the repayment.

Motivation

Since the initial Aave v3 activation three years ago, due to the nature of liquidations, the system has accrued a minimal bad debt of approximately $400’000.

This amount is pretty irrelevant in comparison with the ~$12 billion outstanding borrowings, but given the imminent activation of the Umbrella system (to automatically cover bad debt), and the recent upgrade of Aave to v3.3 (formally accounting for deficit/bad-debt post-liquidation), we believe it is appropriate to raise a proposal to clean the “legacy” bad debt, and this way start with clean accounting.

The Aave ClinicSteward is a smart contract that simply facilitates for a permissioned entity to execute the clean-up on behalf of the DAO, by authorizing a constraint budget from the Aave Collector for exclusive usage on liquidations/repayments of unhealthy positions (liquidatable).

Specification

The Stewards were deployed with the following configuration:

  • the governance executor lvl1 was granted the DEFAULT_ADMIN role on each network
  • a multisig with 2/3 multisig with ACI/TokenLogic/Karpatkey as signers was granted the CLEANUP_ROLE admin

The proposal:

  • grants the ClinicSteward the FUNDS_ADMIN role in the Aave DAO.
  • grants an EOA/Bot controlled by ACI the CLEANUP_ROLE to cleanup debt on behalf of the Aave DAO.

The stewards themselves implement a $ cap, which is pre-configured to the following amounts:

  • Ethereum core $30’000
  • Ethereum prime $5’000
  • Polygon $30’000
  • Avalanche $350’000
  • Arbitrum $60’000
  • Optimism $5’000
  • Base $15’000
  • Metis $1’000
  • Gnosis Chain $1’000
  • BNB Chain $2’000
  • Scroll $1’000
  • ZKSync $1’000
  • Linea $1’000

References

Copyright

Copyright and related rights waived via CC0.

03Mismatch analysis

AIMismatch analysis by Kvorumlow confidenceclaude-sonnet-5
Consistent

The proposal follows Aave's standard governance pattern where the top-level on-chain proposal only contains calls to `execute()` on one or more Payload contracts; the actual granular logic (role grants, cap configuration, Collector fund approvals) is implemented inside those payload contracts via delegatecall and is not visible in the decoded top-level calldata. Both decoded actions here are simply fallback `execute()` calls to two addresses, giving no decoded arguments to verify against the specific claims in the description (FUNDS_ADMIN grant, CLEANUP_ROLE grant to an ACI EOA, per-network USD caps, DEFAULT_ADMIN role assignment). Because the calldata does not expose enough detail to confirm or contradict any of these specific claims, I cannot assert a match or a mismatch with confidence. Nothing in the visible calldata contradicts the description, and the pattern (two payload execute() calls, one likely per group of networks) is consistent with a multi-chain steward activation as described. Given the opacity of the decoded actions, confidence is low, but there is no evidence of any inserted extra action, target mismatch, or omitted step that a reviewer could point to from what's decoded. This is a common limitation in reviewing Aave payload-based proposals, and does not itself indicate malicious or careless drafting, so I record the assessment as consistent with low confidence rather than flagging speculative discrepancies.

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
0x4cab…c0fc
Value
0 wei
Function
fallback()
Arguments
{}
Calldata
#1fallback()chain 0x1
Target
0x2990…7fa9
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.