Skip to content
PROTOCOL RESEARCH

Rysk risk

Clean covered-call mechanics (fully collateralized, physically settled per docs, real premium yield) but upgradeable custody under a 3-of-5 multisig with no timelock at any level of the verified upgrade path, audit reports that would not open even though Rysk's own site names Trust Security, unnamed RFQ buyers, and a hard lock until expiry gate it to caution.
CautionResearched Jul 18, 2026

Scorecard

watch: Audits & contractsRysk's own site claims 5 audits by Trust Security but no report file opened this run; the docs security page lists six rounds with two marked upcoming and no firm named; contracts fork Opyn Gamma and the collateral pool is an upgradeable proxy.
watch: Admin controlVerified end to end on-chain: both MarginPools chain through an AddressBook-style contract to the same 3-of-5 Gnosis Safe on Ethereum and HyperEVM, with no timelock at the proxy, AddressBook, or Safe level and unidentified signers.
watch: OracleStork primary with a Pyth divergence halt is reasonable design, but Stork is young and a manual Rysk team check sits inside the 5-minute dispute window at settlement.
weak: Liquidity & exitNo exit before expiry at all; the FAQ says a secondary market is not yet built, so a 6-week call is a 6-week hard lock with no roll or cut.
strong: Yield (real vs emission)Premium paid upfront in USD by option buyers, no emissions; live earn table shows UETH covered calls at roughly 4% to 77% APR by strike, with deep OTM strikes near the bottom of the range.
watch: Holder concentrationNo liquid token (points only) and depositor concentration is unmeasured; weekly TVL swings around expiries look like settlement mechanics rather than whale evidence either way.
watch: Track recordNo exploit found in hacks lists or news, but V12 is only about 13 months old, V1 sits in withdraw-only mode, and TVL is down 44% from the March ATH.
🟢 strong🟡 watch / caveat🔴 weak / fund-loss risk
Verdict is a gate (worst flaw wins), not an average. Our read, not financial advice.

Live facts

auto-sourced now
TVL—
30dflat
Auditsn/a
Last hacknone

DeFiLlama + our exploits feed. Cross-check the dated report against today.

TL;DR

Rysk V12 sells covered calls and cash-secured puts through an on-chain fork of the Opyn Gamma contracts plus an off-chain RFQ auction where makers bid for the premium. Collateral sits in an upgradeable MarginPool; this run traced the upgrade path end to end and it tops out at a 3-of-5 Safe with no timelock at any level. Positions are fully collateralized, settled at a Stork oracle price at the 08:00 UTC fixing, and cannot be closed before expiry. The premium yield is real. What is still not verifiable: the audit reports themselves (Rysk's own site names Trust Security, but no report file would open), and who is buying the calls. Caution.

Checklist

Audits & contracts. The docs security page lists six audit rounds (V12 mk0 through mk3 plus a Lean Gamma upgrade audit and Rysk Premium mk0) with no firm named, and two of the six are marked upcoming. The report links redirect to 404 pages, not PDFs. Rysk's own marketing site states it has "completed 5 audits by Trust Security" and points to a reports page that served an app dashboard instead of reports this run. The firm attribution is therefore first-party, but the reports and their scope are unverifiable: refresh. Dedaub audited Rysk V1 in 2022; those audits do not cover V12 and are not carried over. The contracts fork Opyn Gamma, and the pool holding collateral is an OwnedUpgradeabilityProxy (verified source on Etherscan), so the code guarding funds can be swapped.

Admin control. Verified on-chain this run, end to end. The Ethereum MarginPool (0x684404F2...C671) and the HyperEVM MarginPool (0x24a44f1d...91aB4) each have a proxyOwner that is an AddressBook-style contract (0x65852e9c...562e7 on Ethereum, 0xffce2d20...93f34 on HyperEVM). The owner of both AddressBooks, and the owner() of both MarginPools, is the same Gnosis Safe 0xafe32eb8...23db4, threshold 3 of 5, identical five signer addresses on both chains. The Safe has no modules, the AddressBook exposes no delay functions, and the proxy's upgradeTo is gated only by onlyProxyOwner. No timelock anywhere in the path. Signers are addresses only, identities unknown.

Oracle. Stork is primary, cross-checked against Pyth with a halt if they diverge past a per-asset threshold; a few assets (wstHYPE, Pendle assets) price from Pyth or the Pendle API directly. Settlement uses the 08:00 UTC price fixing, with a manual Rysk team check inside a 5-minute contract-enforced dispute period. Sound design on paper, but Stork is a young provider and there is a human in the settlement loop.

Liquidity & exit. None before expiry. The FAQ is explicit: "Not yet. Rysk is developing a secondary market that will allow users to buy back or offset open positions before expiry." A 6-week call means 6 weeks of hard lock. If the underlying runs through the strike mid-cycle, you cannot roll, hedge out, or cut.

Yield: real vs emission. Real. Premiums are paid upfront in USD by option buyers, with no emissions on top. The live earn table this run lists UETH covered calls at roughly 4% to 77% APR depending on strike (WETH itself is not listed; the ETH asset on HyperEVM is UETH), with other assets spanning about 2.5% at the low end (UBTC) to 144% at the top (UPUMP). A 28-30% OTM strike sits near the bottom of its asset's range. The live quote for the exact $2,400 ETH call expiring 2026-08-28: refresh. Do not underwrite at headline max-APR numbers. One flag: app marketing copy describes covered calls as cash-settled while the docs say physically settled with collateral exchanged at strike; which text is stale: refresh.

Holder concentration. No liquid token, points program only. Depositor concentration inside the MarginPool is unmeasured: refresh. TVL swings hard around weekly expiries ($59.9M to $47.2M on 07-10/11, $56.4M to $37.3M on 07-17/18), which looks more like settlement mechanics than a single whale, but nothing here is measured.

Worst case

The 3-of-5 Safe upgrades the MarginPool implementation to a malicious one and every locked position is drained; this run verified there is no timelock at the proxy, AddressBook, or Safe level, so nothing in the upgrade path forces a delay you could react to. Short of that, ETH settles above your strike at the 08:00 fixing and your collateral is exchanged at strike after six weeks in which you could not exit, or a Stork mispricing settles the book at a bad price inside the 5-minute manual dispute window. There is no insurance fund. Position sizing is the only mitigation.

Bottom line

Caution. The mechanics are clean for what they are: full collateralization, physical settlement per docs, premium upfront, no emissions, and the docs state no rehypothecation. What keeps it out of solid is everything around the mechanics: audit reports that would not open even though Rysk's own site names Trust Security, upgradeable custody under a 3-of-5 multisig with no timelock, unnamed RFQ buyers, and a hard lock until expiry. The track record is clean so far: no entry in the DeFiLlama hacks list and a clean news scan, with V12 about 13 months old on HyperEVM and the 2022-2023 Arbitrum V1 sitting in withdraw-only mode with about $193k left. Fits only as a small overlay, and the audit provenance needs to be pinned down before any renewal.

Data appendix

  • TVL: $40.5M (DeFiLlama API, verified this run). Split: HyperEVM $38.1M, Ethereum $2.2M, Arbitrum V1 remnant $0.19M. ATH $72.56M (2026-03-27); 90d ago $57.1M (2026-04-20). Weekly sawtooth: $59.9M to $47.2M (07-10 to 07-11) and $56.4M to $37.3M (07-17 to 07-18, partially recovered to $40.5M intraday); cause refresh (weekly expiry settlement plausible given the repeat pattern).
  • Audits: docs.rysk.finance/resources/security lists six rounds, two marked upcoming, no firm named; report links returned HTML/404, not PDFs, this run. Rysk's own site (ryskfi.com/premium) claims "5 audits by Trust Security"; its reports page served no reports. Report content and scope: refresh. Dedaub audits (Jun 2022, Dec 2022) cover V1 only.
  • Admin/governance: MarginPool proxies on Ethereum (0x684404F2AEBAD87a6803F13741B1d638Bfe2C671) and HyperEVM (0x24a44f1dc25540c62c1196FfC297dFC951C91aB4) are OwnedUpgradeabilityProxy contracts (Etherscan-verified source, Solidity 0.6.10). proxyOwner: AddressBook-style contracts 0x65852e9cf13d1a3f330be2b95b2c1b4396d562e7 (Ethereum) and 0xffce2d20e0f68dcedbce657175684845f9593f34 (HyperEVM). Owner of both, and owner() of both MarginPools: Safe 0xafe32eb89391dfd5900f98857f009477e4423db4, threshold 3 of 5, same five signers on both chains, no Safe modules; all read via eth_call this run. Timelock: none found at any of the three levels.
  • Oracle: Stork primary, Pyth divergence check with system halt, manual team check plus 5-minute contract-enforced dispute period at the 08:00 UTC fixing (docs.rysk.finance/resources/oracle).
  • Settlement: European options, fully collateralized, physically settled per docs; ITM collateral exchanged at strike; payout within about 2 hours of expiry (docs FAQ). App marketing copy says covered calls are cash-settled: discrepancy, refresh.
  • Counterparty: open RFQ, integrated makers bid, "anyone can act as a counterparty" per the FAQ; docs state full collateralization and no rehypothecation; maker identities and on-chain enforcement of the buyer side: refresh.
  • Holder concentration: no token; depositor concentration: refresh.
  • Recent-news scan: no incident found; no Rysk entry in the DeFiLlama hacks list (588 entries, checked 2026-07-18).

Maintained monthly. Methodology: DeFi Research Instruction v2.

← all protocols