Diligence reference, July 2026
A risk checklist for delta-neutral synthetic dollars.
Ten questions to put to any delta-neutral dollar before depositing, each paired with how to check it yourself rather than take the answer on faith. At the end of each item is Kerne's own honest answer, including the places Kerne cannot be checked or has not yet earned the benefit of the doubt.
The delta-neutral synthetic dollars that failed in the last cycle did not fail because the trade is unsound. They failed on rented capital, privileged keys, off-chain collateral, and yields that were advertised but never realized. The checklist below is organized around those failure points, so running it is really asking whether a given dollar repeats a mistake the sector has already made. A fuller account of what actually broke is on the failure ledger and in the synthetic-dollar graveyard.
1.Negative funding
Ask
The yield comes mostly from perpetual funding. What happens to the holder when funding turns negative, and who absorbs it?
How to check it
Look for a dedicated insurance or reserve fund sized against negative-funding episodes, and a published funding regime. A protocol that only shows the good months is hiding the mechanism, not removing it.
Kerne's own answer
Kerne runs a dynamic insurance allocation against negative-funding periods and publishes the live funding regime and a forecast at /api/forecast. The yield is still variable and can go to zero in a negative regime, and Kerne says so.
2.Hedge and venue risk
Ask
Where does the short hedge run, and can you verify it exists and is the size it claims?
How to check it
Ask which venue holds the hedge and whether its size is provable. For almost every delta-neutral dollar the hedge is on a centralized exchange and is not user-verifiable; the best a protocol can do is a signed attestation, which is not proof.
Kerne's own answer
Kerne's hedge runs on Hyperliquid and is self-reported and signature-bound only. It is as unverifiable to a holder as any competitor's. This is stated as a load-bearing limitation on the Honesty Index, not glossed.
3.Collateral custody
Ask
Where does the collateral backing the dollar actually sit, and who can move it?
How to check it
Distinguish on-chain collateral you can read from collateral held by a centralized custodian you must trust. Neither is wrong, but they are different trust assumptions and the protocol should say which it is.
Kerne's own answer
kUSD's peg reserve is USDC held one to one in an on-chain module, and the vault collateral is on-chain and readable on Base. The hedge margin, by contrast, sits on Hyperliquid. Kerne discloses both legs rather than implying everything is on-chain.
4.Privileged mint keys
Ask
Can any single key mint the dollar, and if so, what stops an unbounded mint?
How to check it
Read who holds the mint role. A single externally-owned account with mint authority is the exact failure that let one key mint tens of millions of a stablecoin in a past incident. Prefer mint gated to an on-chain contract with a rule, administered by a multisig.
Kerne's own answer
kUSD's mint role is held by an on-chain module, not a personal key, and role administration is gated by a 2-of-3 Safe. Kerne audited its own keys in public after the single-key incident elsewhere; read that at /insights.
5.Oracle and attestation integrity
Ask
If the protocol publishes a signed price or reserve, is anyone actually checking the signature, and is the signer the one you expect?
How to check it
Recover the signer from a published signature and confirm it matches the address the protocol claims. Two oracle failures in one week came down to a signature nobody verified. A signature you do not check is decoration.
Kerne's own answer
Kerne signs its Proof of Reserves and its Honesty Index snapshots, publishes the signer address, and shipped, then disclosed, a bug in its own May signature check. The method to verify any signature is at /verify.
6.Redemption and exit path
Ask
How does a holder get back to a hard dollar, at what ratio, and can that path be paused?
How to check it
Find the redeem mechanism and whether it is one to one against a reserve you can see, or a market sale that can gap in a rush. Check whether redemption can be halted, and under what authority.
Kerne's own answer
kUSD redeems against an on-chain USDC reserve; the reserve and the outstanding supply are both readable, so the coverage is checkable at any block. Kerne is small, so exit-liquidity depth is limited, which is a real constraint at size and is stated.
7.Reserve transparency
Ask
Can you recompute the reserves yourself, or must you trust an attestor's periodic report?
How to check it
Separate recomputation, where you read the chain, from attestation, where a firm certifies a number at a point in time. Both are legitimate; only one removes the attestor from the trust path. This axis is scored on its own scorecard.
Kerne's own answer
Kerne's reserve leg is recomputable from on-chain reads listed in its PoR endpoint. The hedge leg is not; it is attested. So Kerne is recomputable on part of the book and attested on the rest, and the scorecard says exactly that.
8.Advertised versus realized yield
Ask
Has the protocol actually paid what it advertises, and can you tell?
How to check it
Read the staked token's ERC-4626 share price at two blocks and annualize the growth. Compare it against the advertised rate over the same window and basis. A modeled or since-inception number is not a record of what holders earned.
Kerne's own answer
On Kerne's own Honesty Index, skUSD advertises a modeled APY and has a realized figure near zero, the widest gap on the board. Kerne has not yet paid meaningful realized yield, and lists its own row first.
9.Smart contract and audit status
Ask
Is the code audited, and does the deployed bytecode match the audited source?
How to check it
Check for a public audit report and a source-verification status on a block explorer, and whether the protocol discloses any place live bytecode differs from current source.
Kerne's own answer
Kerne is pre-audit: Hexens fieldwork is underway with no public report yet. It publishes a deployed-versus-source disclosure at /security/deployed-vs-source and a findings-response protocol. Until the report lands, treat the code as unaudited.
10.Underlying peg
Ask
Is the dollar's realized yield denominated in something that is itself at a dollar?
How to check it
A staked token's share price is measured in its underlying. If the underlying has depegged, steady share-price growth can hide a dollar loss. Check the underlying's own price, not just the vault's growth.
Kerne's own answer
kUSD is minted one to one from USDC and its peg reserve is USDC, so its underlying is a hard dollar rather than a floating asset. The failure this item warns about, a vault growing in a broken underlying, is one Kerne documents in other protocols on /legible.
The honest tally for Kerne.
Run this checklist against Kerne and it passes the custody, mint-key, peg, and reserve-recompute questions, and it fails or comes up short on three: it is pre-audit, its hedge leg is not user-verifiable, and it has not yet paid meaningful realized yield. Those are the same three things Kerne lists against itself on the Honesty Index. A checklist that only made a protocol look good would not be worth running, including this one.
Related reading.
Verify Kerne against this list yourself: the transparency dashboard, /api/por, /api/risk-status, and the machine-readable /facts.json. The yield-honesty axis is the Honesty Index; the reserve-transparency axis is the scorecard; the trust models are mapped at attestation vs recomputation.
Also: advertised vs realized APY, skUSD vs sUSDe vs sNUSD, and stablecoin proof of reserves on Base.
This is a diligence aid, not a rating and not financial advice. Every protocol named other than Kerne is referenced for comparison only, from its own public disclosures.