Researcher | ex @chaoslabs
Joined February 2023
- Tweets 1,805
- Following 1,000
- Followers 2,287
- Likes 12,290
210 Photos and videos
Jun 3
Sources: I relied heavily (but not exclusively) on this excellent article by Justin Thaler (@SuccinctJT):
a16zcrypto.com/posts/article…
Image source (CC BY-NC 2.0): flickr.com/photos/feuilllu/4…
On how Ethereum is tackling the risks: pq.ethereum.org/
138
Jun 2
My read of this is that IF we can build reliable prediction-market oracles (which seems like a much easier problem than building real time oracles) THEN we can unlock this.
Risk perspective: "liquidation hunting" will be replaced with "rollover hunting"
x.com/VitalikButerin/status/…
Jun 1
Building index-tracking assets on top of options instead of debt
ethresear.ch/t/building-inde…
What if the use options as the base of defi, instead of CDPs and liquidations? So instead of extreme price movements creating a sharp and global "you get liquidated" effect, instead your exposure to the index diverges quadratically from your preferred exposure in a smoother way?
A key benefit is getting rid of the need for instant oracles, and instead making everything work on top of "slow oracles" (ie. the type that prediction markets use)
This design has a significant downside - the need to do regular rebalancing - and an open question of whether and how this rebalancing can be made slippage-resistant enough. But it's worth considering and trying IMO. I would feel much safer holding algostables inside something like this, than in something that depends on an oracle that has to give real-time answers (and therefore could be tricked into giving wrong real-time answers with no time for human recourse).
7
286
atiselsts.eth retweeted
Apr 15
For one, I believe that audits (for code bugs) are one of the things that DeFi does get right.
And better audits or formal verification wouldn't have prevented most of the recent exploits:
Resolv, Drift > private key compromise lack to timelock and other circuit breakers
Venus Protocol > economic attack, flagged by audits but ignored
Aperture Finance, Solv > bugs in unaudited contracts
Makina Finance > vault logic compromise, OOS for audits
In TradFi, companies are mandated to do thorough risk management by law. In DeFi, we're not. Solution? More social consensus that DeFi needs to learn to take better care of its risks. This already works with code audits - everyone believes that a protocol that doesn't have code reviewed by security experts can't be taken seriously. It also works for L2s because we have @l2beat.
Comparably, we don't have nearly as good insight into the admin control levels of major DeFi protocol deployments. Similarly, we know far more about how to exit L2s in times of crisis than how to exit major DeFi protocols if/when their admin decides to do something funky.
EF doesn't need to pick and choose some favorite protocols. But it would be nice to recognize that there's a crisis in "low-risk" DeFi and push for a more principled approach, where risk management goes beyond looking for bugs in the code.
2
1
11
1,263
Apr 13
mostly two reasons:
- realization that LVR is only small part of what makes onchain liquidity challenging (vindication for us who said it all along)
- realization that AMMs are typically money sinks, not money generators (vindication for OB folks)
x.com/duelinggalois/status/2…
Do you know why you don't hear about LVR anymore? Because everyone lost.
4
2
12
1,906
Apr 7
Low risk DeFi can't be the revenue engine for @ethereum if low risk DeFi does not exist. (Other than a couple of OG protocols like Uniswap)
EF is already doing plenty to make the core protocol more secure. Next step is including DeFi too, like Solana has x.com/SolanaFndn/status/2041…
Apr 6
Solana was built for security. As the ecosystem scales, so does our investment in the tools, standards, and support.
Today that commitment deepens with a new security program, active monitoring, formal verification for top protocols, and a new crisis response network.
Learn more 👇
6
15
2,631
Apr 4
Uniswap UI will finally pick up new hooks automatically.
This used to be a big friction point for teams looking to build on top of v4. Hooks were permissionless, but getting orderflow was not. Now most new hooks will be autorouted. x.com/Uniswap/status/2040114…
Today we're launching hook auto-routing on the Uniswap Web App, Wallet, and API
That means no more allowlist process
Build your hook, deploy your pool, get distribution from day one
4
14
882
atiselsts.eth retweeted
Apr 1
1/ Drift's admin key was compromised.
$213M drained from @solana's largest DEX in under 10 seconds.
Unfortunately, we've seen similar patterns before:
- fake collateral market
- a manipulated oracle
- disabled circuit breakers
Let's break it down 👇
written w/ Chaos AI
42
99
452
174,707
Mar 25
Shows that protocols should be designed in a way that the protocol can be forced to enter a safe state, like:
- paused
- functions selectively disabled
- liquidity pulled from DEX (limits the exit)
The hard question is who's permitted to do this. There's a tradeoff between time-to-react and centralization risks.
But how about this idea: user-controlled pause. If anyone submits a zk proof that a key invariant can be broken → protocol auto-pauses pays a bounty.
x.com/pashov/status/20367430…
Web3 Security Horror Story Time
A protocol gets reported a Critical vulnerability. They immediately patch it with a code fix and push it on-chain to their upgradeable contracts.
A MEV bot picks up the "code fix" transaction before it is validated into a block, re-engineers the vulnerability with AI and front-runs the upgrade patch with an exploit.
Upgrade passes successfully, the exploit before it as well. You just exposed the fix of a Critical vulnerability to an untrusted actor. AI allowed seconds to be enough to deduct a vulnerability from a patch.
You can argue AI is dumb, sure. But you can't argue AI is not fast - and that it can't be even faster. Upgradeability and MEV bots become an attack vector with time.
I challenge you to say how this can be safely secured.
8
610
Mar 23
It took just one compromised EOA to mint $80M USR, but 3 hours to gather signatures to pause @resolvprotocol, and hours more for integrators to react.
1
7
1,160
Mar 23
Analysis of the fallout:
x.com/omeragoldberg/status/2…
Mar 22
1/ Millions in bad debt, at the time of writing, were created across Gauntlet's Morpho vaults from the Resolv USR exploit.
Almost all of it was supplied ** after ** the exploit.
So why would curators supply millions in USDC to a broken market?
Let’s dive in.
1
2
218
Mar 23
On pausing the protocol:
x.com/yieldsandmore/status/2…
Mar 22
It took @ResolvLabs 3h to pause their protocol. Roughly one hour of that delay came from the gap between submitting the multisig transaction and collecting the 4 required signatures to execute it.
We should learn from this. Emergency pausing should only require a SINGLE signature, with that authority distributed across as many of your team members as possible.
Including trusted external operators (like us) also makes sense. You'll get more eyes on any abnormalities onchain, a higher likelihood of a quick pause, and a better coverage of timezones.
Teams should also have reliable communication channels to be coordinate instantly during incidents like this. Pagers, phones, whatever works.
Tools like @HypernativeLabs, @hexagate_ and others offer proactive monitoring that can pause the protocol for you, and could one day save you.
2
179
atiselsts.eth retweeted
Mar 16
Most yield products optimize for APY and bolt on risk
as an afterthought.
All @chaoslabs vaults are powered by the same Chaos AI risk models that have secured trillions in vol and billions in asset deposits.
The appetite for Low-Risk DeFi is real.
Higher.
Mar 16
UPDATE: $35M deposited into Chaos @Veda_labs Vaults on @Krakenfx DeFi Earn & @Krak
• Balanced & Boosted strategies, risk-first by design and powered by the Chaos intelligence suite
• Real-time monitoring across solvency, liquidity, and yield volatility
• Instant redemptions
1
9
42
4,827
Mar 16
Correction: multiple @CoWSwap solvers submitted quotes, but just one could be executed.
And if you're wondering what's the point of a "minSolvers" flag since the user can set max slippage: info asymmetry. Solvers know the fair price better than the user.
x.com/atiselsts_eth/status/2…
Mar 13
In addition to everything else, @CoWSwap orders should be able to specify: "only execute if ≥ N solvers participated." Here just one solver submitted a solution. There was no competitive auction.
x.com/Ehsan1579/status/20323…
2
1
7
1,180
Mar 13
In addition to everything else, @CoWSwap orders should be able to specify: "only execute if ≥ N solvers participated." Here just one solver submitted a solution. There was no competitive auction.
x.com/Ehsan1579/status/20323…
7
1
48
9,704
Mar 11
Gauntlet decided to cancel the proposal, it's now gone from Snapshot.
Shows that the Uni DAO still has some power left!
Broadly speaking, this is what DAOs should do: check and call out BS, instead of trying to execute on their own.
x.com/atiselsts_eth/status/2…
1
4
62
5,869
atiselsts.eth retweeted
Mar 10
1/ stETH CAPO Misconfiguration
Today, a misconfiguration on Aave's CAPO oracle caused wstETH E-Mode liquidations, resulting in a loss of 345 ETH.
No bad debt was incurred, and all affected users will be fully reimbursed.
More below.
24
43
174
74,062
Mar 10
4
636