Vaulture

Our project is named Vaulture. The idea is to create 3 ERC 4626 vaults with different yield strategies.

First vault: We have implemented a first strategy of leverage consisting in the user selecting that vault. Upon selection, the user deposits ETH. Then, through the smart contract, that ETH is getting swapped to ezETH which gets deposited as collateral on Silo. After, borrowing the smart contract borrows WETH, swaps that WETH back to ezETH using 1Inch. We deposit once again the swapped ezETH back on Silo and start over in order to leverage the position through looping in the vault. The LTV (loan-to-value) ratio limits the amount borrowable. The total deposit APR and the total borrow APR let us calculate how much the operation will cost, minus the swapping fees on 1inch or other Dexes. By doing so, we accumulate Eigen Layer points and Renzo Points. If we were able to do so on Mode using Sturdy, we would have been able to accrue Mode, Ether.fi and Eigen Layer points. We were unfortunately not able to implement that strategy. By calculating the accruing point rate of the user to the total accruing point rate of the considered cryptocurrency, and by multiplying that by the share allocated to the community for the considered airdrop season, we can approximate the amount of tokens the user will receive. Then, with the current price of the token for seasons > 1, or with some approximations using for example Whales Market's secondary point market, we can determine the APY of the user. This strategy works because of the ezETH to WETH/ETH peg. As seen before the first airdrop of Renzo, if people start massively selling ezETH, this can depeg ezETH to WETH and liquidate the leveraged position. It is possible to rebuy the position faster than when the looping was done by redepositing almost all the WETH borrowed to take the collateral back and close the leveraged position to limit the risks of losing money when a depeg is detected.

Second vault: https://github.com/DavidDrob/silo-weth-leverage The idea of the second vault is to let an admin decide to which other ERC 4626 he would like to invest in. So our vault is connected to several different other vaults. According to the assets in the other vaults, we calculate a fear and greed index, using AI sentiment analysis on Reddit/Twitter, to give an idea of the risk associated to invest in such or such vault and the possible rewards. We also allow staking and liquidity-providing to generate yield in these different vaults.

Third vault: This one is a more flexible option. One would be a kickstarter vault. The user could deposit USDC to invest in a food truck, that USDC is withdrawn, food is bought and sold, and then after a week the money made is deposited back into the contract with the profits. Another one, onchain, would be a user choosing their own pools on aerodrome or some strategy they will do manually rather than codify the exact strategy.

Vault 1: We utilized the YearnV3 Vault template to develop a strategy that leverages wrapped ETH (wETH) as the want token, interacting with the Silo.finance protocol on the Optimism Mainnet.

Upon user deposit, the strategy swaps wETH for ezETH, which is then leveraged three times to maximize the fund's potential while minimizing the risk of liquidation. The ezETH is subsequently deposited into Silo, where it generates multiple rewards:

• EigenLayer points
• 3x Renzo points
• SuperFest (OP) rewards
• Deposit APR (4.6% at the time of writing)

The strategy can be harvested at any time. During this process, the borrowed amount is repaid, and the initial funds, which are expected to have grown, are withdrawn back into the strategy contract. The deposit process is then repeated, with a new deposit utilizing the 3x leverage.

Since this strategy operates within an ERC-4626 Vault, users have the flexibility to withdraw their funds from the vault at any time.

Vault 2: https://github.com/DavidDrob/silo-weth-leverage For the sentiment analysis part, several options have been explored, with no results unfortunately because of technical difficulties of products not being compatible with current version of websites or pratical difficulties being projects not letting us tests their products for free, or products not having the relevant cryptos and only covering bitcoin and/or ethereum. Amongst the solutions tested to scrape Twitter or directly get sentiment analysis that have not worked for us:

• snscrape
• stockgeist.ai
• scrapingdog
• Apify
• surgehq.ai
• alternative.me
• crypto-sentiment.com
• coincodex's sentiment analyzer
• sanbase
• senticrypt
• rishikonapure's Cryptocurrency-Sentiment-Analysis

In the end, we were able to obtain a temporary key from augmento.ai (thanks again to them!). The data provided reflects already sentiment-analyzed tweets on 77 crypto assets. So we aggregated the data to only 3 categories, positive, neutral and negative, in order to use that data in our home made Fear & Greed (F&G) index. The formula of the F&G being the following:

0.30 * volatility + 0.30 * momentum_volume + 0.15 * dominance + 0.25 * crypto sentiment

Unfortunately, we were not able to get the numbers for the volatility and the momentum_volume from CoinMarketCap nor CoinGecko since these data are accessible through paid API.

The aim of this part of the project is to give an easy insight to the user to decide with which strategy to go for, balancing risk and APY.

All this work has been done in Python.

Third Vault Strategy: In order to allow the funds to be used off-chain but still be composable in 4626 vaults we had to create an implementation that would transfer the funds to an arbitrary address that could manually re-deposit the asset after either a profit or loss scenario.

This is an anti-pattern of sorts as the “freeFunds” function is basically proved ineffective, perhaps only as a way of alerting the off-chain keeper that he needs to deposit some asset or liquidize it fully depending on the strategy.

The underlying “report” functionality needs to be updated above and beyond the _harvestAndReport function as it needs to be self aware of how much of the asset is being returned and trigger that transfer itself.

To develop the project, we used a Scaffold-eth fork for foundry and Tanderly to begin with. The development of the first vault solution (ezETH/WETH) has been done on a fork of Optimism mainnet.