[crvUSD]: GHO Pegkeeper Review

Proposals
1

⁠⁠Summary

This proposal adds the GHO/crvUSD pool 0x635EF0056A597D13863B73825CcA297236578595 to the crvUSD PegKeeper set, with an initial 3,000,000 crvUSD debt ceiling. The goal is to diversify PegKeeper counterparties while keeping initial exposure conservative, given the pool’s current depth, and to enable a monitored, staged ramp-up as liquidity matures.

This proposal is based on the accompanying Pegkeeper Onboarding Review: Aave’s GHO, co-authored by TokenLogic and LlamaRisk.

Motivation

image
image1600×900 107 KB

Why GHO?

  • Meaningful on-chain peg infrastructure: GHO’s Stability Module (GSM) provides a strong on-chain conversion rail (USDC/USDT ↔ GHO) and has historically supported tight peg behavior within a relatively stable band.
  • Diversification of PegKeeper basket: Adding GHO expands the PegKeeper set beyond the existing mix (USDC/USDT/PYUSD/frxUSD), improving resilience to idiosyncratic risks.

Why conservative sizing?

  • The GHO/crvUSD pool is currently small relative to incumbent PegKeeper pools (per the report), so a small ceiling is appropriate to avoid PegKeeper actions becoming prone to noise.

Proposed Parameters

Initial debt ceiling: 3,000,000 crvUSD

Ramping principle (non-binding guidance for future votes):

  • Consider increasing to 10–15m only once the GHO/crvUSD pool sustains a meaningful share of total PegKeeper TVL (e.g., low-single-digit %, ~5% heuristic of overall pegkeeper), with stable execution quality.
  • Consider ceilings >15m only if system-level peg assurance remains strong, with explicit attention to GSM depth/usability and governance changes that could widen effective redemption bands.

Risk Considerations (summarized)

Key risks are not “tail depeg” style historically, but rather:

  • Structural peg width: GHO can trade in a modestly wider “normal band” than the tightest assets (e.g., USDC). This is acceptable if exposure is sized appropriately and PegKeeper safeguards remain active.
  • Pool depth/reflexivity risk: If PegKeeper debt is large relative to pool depth, the PegKeeper can become the dominant flow, increasing slippage and potentially amplifying short-term deviations.
  • Governance/parameter risk: GSM fees/caps and broader governance decisions can affect effective arbitrage bounds and liquidity.

Mitigants:

  • Curve’s PegKeeper Regulator can effectively halt PegKeeper action by setting allowed amounts to zero if checks fail.
  • Starting with 3m keeps the worst-case impact bounded while the pool grows.

Disclaimer

TokenLogic and LlamaRisk are both service providers to Aave. Neither have been commissioned or compensated specifically for producing this report. This proposal and the underlying report are provided for informational purposes and should not be construed as legal, financial, tax, or other professional advice.

1 Like
2

A GHO PegKeeper and price adapter were deployed:
Pegkeeper: https://etherscan.io/address/0x53876b157decf04389eed66c7c29d73863f8c50b
Adapter: https://etherscan.io/address/0x697fc22314e736056eb0ad7150167ccc9d3396db

Vote has been initiated with payload:

POOL_NG_PROXY = "0x697fc22314e736056eb0ad7150167ccc9d3396db"
PK = "0x53876b157decf04389eed66c7c29d73863f8c50b"
FACTORY = "0xC9332fdCB1C491Dcc683bAe86Fe3cb70360738BC"
REGULATOR = "0x36a04CAffc681fa179558B2Aaba30395CDdd855f"
OWNER_PROXY = "0xb7400D2EA0f6DC1d7b153aA430B9E572F28afB79"

debt_ceiling = 3_000_000 * 10 ** 18

MONETARY_POLICIES = [
    "0x07491D124ddB3Ef59a8938fCB3EE50F9FA0b9251",  # single MonPol
]

AGGREGATE_ORACLES = [
    "0x18672b1b0c623a30089A280Ed9256379fb0E4E62",
]

factory = Contract(FACTORY)
pk = web3.toChecksumAddress(PK)
contract = web3.eth.contract(address=FACTORY, abi=factory.abi)
calldata = contract.encodeABI("set_debt_ceiling",[pk, debt_ceiling])

ACTIONS = [
    (OWNER_PROXY, "execute", FACTORY, calldata),
    (REGULATOR, "add_peg_keepers", [PK]),
    *[(m_policy, "add_peg_keeper", PK) for m_policy in MONETARY_POLICIES],
    *[(oracle, "add_price_pair", POOL_NG_PROXY) for oracle in AGGREGATE_ORACLES],
]

Vote here: