Kashi - Add permissionless oracles through the Open Oracle Framework

Oracle feeds are hard to create and establish, especially for smaller projects, which stops open access and creates permissioned access to otherwise permissionless tools. Chainlink are a bitch to get new feeds for. So we (Conjure) made the Open Oracle Framework, which allows even basic devs to deploy a feed for any asset (technically any api) within 20m for completely new, from scratch, no knowledge about the platform feeds, allowing for any tokens to actually have markets (not just those with link feeds).

Allow support for Open Oracle Framework feeds to let developers deploy oracles to establish markets for any token asset for Kashi for lending, permissionless.

Chainlink sucks for deploying feeds. OOF doesn’t suck and can be done fast and easily.

Allow for Kashis oracle feeds to take Open Oracle Framework feeds. Can use our raw oracle feeds or the TWAP constructor.


Interfaces ready to use

Support OOF for Kashi as an oracle able to be used for new pools

Don’t support OOF

Unable to create a poll


Hi Pr0,

Thanks for your proposal - my initial thoughts against this proposal is that using “any asset” with “technically any API” could lead to centralized and/or low-quality data being used on Kashi. This puts Sushi and its users at risk to oracle attacks and related exploits. The Chainlink price feeds that are currently supporting Kashi are the most battle-tested oracles in the DeFi ecosystem, already securing billions of dollars. I cannot find which protocols that OOF is supporting with their price feeds if you can provide any more info on that to check out the track record.

Any reasoning for it being difficult to add new price feeds to Chainlink oracles is because of the rigorous standards for variably-sourced, high quality data with node operators maintaining consistent uptime and reliability. Price feeds are added by user demand and the amount of feeds are growing constantly, with the current list on Kashi being fairly extensive. Allowing “any tokens” (especially those with smaller markets/liquidity) can really be dangerous because it opens up the door to price manipulation that can part users with their funds when using Kashi since it hasn’t gone through the same security measures.


I agree. The additional assets we’d gain don’t outweigh the risk we’d take on from an oracle solution that isn’t battle tested. There is a reason Chainlink has quality control standards/measures in place when considering new feed additions, which is why they are trusted. If users were to lose funds via manipulation of an OOF oracle…then faith in the entire system will be compromised.

cc: @Clearwood

I think generally speaking we want to support more Oracles just need to make sure they are reliable and easy to integrate added Clearwood for context and a poll for people to vote on this topic.

Another question is who pays for them? If there is no cost associated or minimal for the end-users or Sushi I don’t see the problem.

People should be able to create their own Kashi market with the Oracle they want without us deciding for them.

  • Support OOF for Kashi as an oracle able to be used for new pools
  • Don’t support OOF

0 voters

Personally agree with open permissionless and free market models. Noone assumes that sushi or Uniswap are vulnerable when people use them to sell rugpulls. But both markets have grown through free and open access. The operational costs are paid by the oracle provider. Some projects may run these for public good (have personally spent 15 ETH for public feeds) while others would probably be run by the projects who run the protocol for the asset which the markets being made for, so they would do so for more use for the tokens.

The pools are not collective security, each pool for lending only has exposure to the own pools oracles, so you can have some pools that are secure and some that are not. Projects running own nodes probably wont tank the whole token for a single lending market for that token.

The real value heres that we have not made a system thats built restrictedand tryd to be secure and then tryd to open up, we have built to be open and allowed oracle operators to choose their own security and then users to have an oracle market. Multisig support starts the foundation but because we are on chajn the oracle signers can just be other contracts pulling values from dexes and other oracles.

Hey @Pr0 if you are available, feel free to come on the Forum Call today at 21:00 UTC and discuss your proposal further!


Unfortunately have been traveling and cofounders at Dubai so was too late. Will hop on when able

Feel free to join our next Forum call! They happen every Thursday at 21:00 UTC.