That would not be a problem at all. In addition, we are actively working on a zk succinct proof based solution that will further enhance security for use cases like governance.
What is the status on the temperature check? we should have a vote as soon as possible. I will be delegating to GFXlabs. Bridges discussion can be held seperate.
I plan to do a temperature check this week!
Thank you @ilia_0x for this proposal.
Speaking on behalf of BNB Chain foundation side, we are welcoming this idea and are humbled by this proposal.
We will also be happy to provide all the necessary support for a successful deployment on the chain if the vote pass.
Does anyone think about the risks?
We’ve been doing a bit of research on the testnet deployment. Below is a list of the immediately relevant contracts:
- BSC Testnet Uniswap V3 Factory contract: link
- BSC Testnet Message Receiver Adapter: link
- BSC Testnet Message Bus: link
- Goerli Message Bus: link
Test use of the Celer Messaging system:
At a high level, we can see the Uniswap v3 protocol has been deployed on testnet and that Celer successfully passed a message from an EOA on Goerli, which implemented an additional fee tier on the BSC testnet deployment.
Overall, we’re pleased with the work that Ilia and his team have performed.
However, in reading through Celer’s deployed contracts and documentation, we have a few concerns and questions we would like addressed.
The Message Bus contracts on Ethereum and BSC mainnet have an owner role. The owner controls access to the following functions: setFeePerByte, setFeeBase, setLiquidityBridge, setPegBridge, setPegVault, setPegBridgeV2, setPegVaultV2, and transfer ownership. In addition to this access, since the contract is upgradable, the owner is able to upgrade the implementation of the contract, which means anything is possible. The owner on the Message Bus contracts is the “SimpleGovernance” contract. However, the governance contract functions akin to a multisig since it has five voters with equal voting power. We were unable to find information regarding the five EOAs on the contract, however they appear to be active.
After reading through the contracts, we tried to find more information on how their “Optimistic-rollup-style” security model works. We were able to find this blog post & doc page, but it only had the same information from the forum post. We did find this reference to a “DelayedTransfer”, but it is unclear how this is “optimistic-like” rather than a simple delay.
We were also unable to find documentation or implementation instructions for running an app guardian. If the Celer team could please share the technical documentation and implementation of their proposed security model, we would appreciate it.
While the messaging contract appears to have been audited, their audits are from PeckShield & Slow Mist which do not inspire a high degree of confidence. Both audits were conducted in February 2022, before the addition of the delay mechanism.
We continue to be supportive of a Uniswap v3 deployment on BSC; however, the Celer team controlling ownership of the Message Bus and limited information on the security model makes it hard for us to support the proposal.
Thank you for the comments and questions. I wanted to reply here first to acknowledge the receipt of your comments. However, since today is the Lunar New Year and many of our team is on family or religious holidays, a detailed reply with technical documentation updates might be a bit delayed. Apologies for this and we will reply as soon as we can.
I explained above that since BNB Chain doesn’t have a canonical bridge, any interoperability solution will inevitably have a validation layer and Uniswap governance will still have to bear the risk that participants of the validation layer will not collude.
zk, optimistic — are just different implementations of the validation layer. As an example, here is an article from L2Beat with an interesting experiment that shows that “oracle and relayer” implementation of the validation layer is a 2/2 consensus mechanism, where validation of headers/proofs doesn’t add any security guarantees, but only gas costs since oracle and relayer can always collude.
I think the best way to move forward for Uniswap is to have a framework for bridges so that instead of relying on a single solution, Governance decisions are sent from Ethereum to other chains (e.g. BNB Chain), my multiple bridges. The passed decision is executed on the destination chain only in case the content of the messages sent by all bridges participating in the framework is matching.
deBridge team can help to prepare this framework so that Uniswap governance can have an established process for adding any new chains in the future
GFX Labs team, thank you again for your questions. We would love to provide some responses to your questions.
On audits and security track records
First, we would like to elaborate a bit on our security audits and track records. We respect that different people may have different opinions about security firms, but we believe that both Peckshield and Slowmist are well-regarded security audit firms. Peckshield, in particular, has been consistently the first to report many security incidents and coordinate major security investigations with top security researchers and firms such as banteg, samczsun, and others. If you have any specific concerns about the quality of these firms, we would be happy to relay them and invite these firms to answer or clarify.
Regarding the audit dates, we built Celer’s system in such a way that our optimistic rollup-like security model can be securely and easily added to dApps as a design pattern. The capability to do so has been available before Feb 2022, and we only began talking about it proactively later when we realized that some teams were marketing it as a unique feature.
In addition to security audits, Celer also has a standing $2M bug bounty on ImmuneFi since November 2021, which has not been claimed yet.
In terms of our security track record, Celer’s cross-chain system is the ONLY ONE that has processed a meaningfully large volume (bigger than $1 billion in asset bridging volume) without any critical security issues being reported or exploited. Additionally, throughout Celer’s development history since 2018, none of the systems or products built by Celer, including the world’s first Generalized State Channel Network, have been exploited in any way. Celer has a security operation team that is active 24/7 to monitor the public blockchain transactions for any anomalies related to Celer’s systems and respond to any reports from security firms and independent security researchers with a 3-minute SLA. Of course, we know blockchain is a dark forest and our flawless record does not automatically guarantee future security, but we hope that the information provided gives you more confidence from a comparative perspective.
On optimistic rollup-like security model
The security model of the optimistic rollup-like system has two components: 1) a two-phase commit pattern in the smart contract, and 2) an off-chain App Guardian that can halt the execution of a message during the delay between commit and confirmation.
We would be happy to explain this in more detail using the DelayedTransfer example you mentioned. DelayedTransfer is related to the optimistic-rollup-like security model used in cBridge, an asset bridge built on the Celer Network. When an asset bridge transaction is initiated and the message is relayed to the destination chain, the message is placed in a DelayedTransfer queue on the destination chain, where it must wait a set delay before funds are released with a confirm transaction.
During this delay, an App Guardian cross-checks the source chain to verify the presence of a corresponding bridge transaction. If a match is found, the App Guardian takes no action and the confirm transaction is executed as normal. However, if a matching transaction is not found (indicating a problem with the bridge), the App Guardian can halt the asset bridge smart contract, preventing any malicious transactions from executing and safeguarding funds. The App Guardian is operated by every validator of SGN and can also be run by additional parties, such as the Uniswap foundation, for their own dApps.
The above security model will be applied to Uniswap’s governance use case. We hope this explanation has provided clarity, but we are available for further discussion via a call or in a Twitter space.
We have recently taken down some of the documentation and reference implementation for the dApp side of the optimistic-rollup-like security model, in order to make it more adaptable for various application use cases. We will follow up with links to the updated documentation and App Guardian repositories within the next two days.
On governance multisig
You are correct in noting that the MessageBus contract is upgradeable through a governance multisig. Our governance process involves a community voting process, and the execution of governance decisions for on-chain contract upgrades is carried out by a subset of validators including InfStone, Binance Staking, OKEX and Celer Foundation.
We made the MessageBus upgradeable with the goal of making it easier to address any potential security issues just in case and add must-have features. However, we approach this process with care and continually evaluate and improve our governance process. We welcome additional active contributors such as GFXLabs to be more involved.
We hope the above clarifies the concerns to some degree and we are always happy to answer any further questions!
Hi Mo, I love the tech that you’re building at Celer, but I think that the Uniswap community may misinterpret a number of your points, so these need to be clarified.
When you say a “Cosmos-consensus Security Model” or “L1-PoS-blockchain security mode,” someone may interpret that Celer inherits the security model from the Cosmos chain or from the layer-1 where the message was initiated — both of these statements are not the case. Instead, it is accurate to say that with this model, the Celer network is using Tendermint and Cosmos SDK as a set of ready-to-use tools and technologies for implementing consensus and networking layers for the project.
The PoS consensus implementation allows anyone with ⅔ of the voting power (i.e. ⅔ of the amount of staked tokens) to have full control over the consensus and validation of cross-chain messages.
The current total stake is 2,587,158,380 CELR and thus, if nodes with more than 2/3 of the total stake act maliciously, they can forge any message. 7 nodes currently have enough stake to control consensus.
⅔ of the total current stake is 1,724,772,247.33 CELR ≈ $25M at current market price. This is the approximate cost that would allow an attacker to withdraw liquidity from all Celer liquidity pools and forge any message to extract value from any Apps built using Celer Network, and, importantly, I assume the potential value extracted by an attack on Uniswap may be higher than the cost of attack.
The only staked asset is the CELR governance token, which means that validators’ collateral directly correlates with the protocol’s performance. This is why an attacker could open a large short perpetual futures position before attacking the protocol, and recover/hedge the cost of their acquired stake.
Having more than 2/3 of the protocol voting power would allow a malicious entity to control the consensus process and make arbitrary decisions about which blocks to add to the blockchain, forge cross-chain messages, reverse and censor transactions, create fake transactions, and create their own set of validation nodes.
Correct me if I’m wrong, but this attack can be performed by any single whitelisted Celer validator now.
Optimistic security model
You may say that the described problem can be solved through an Optimistic rollup-like security model which is the second alternative provided by Celer. My guess is that this wouldn’t add any additional security to the validation layer for a few reasons:
- The optimistic component can always be implemented on the application level, where the receiving smart contract in BNB Chain can have a “quarantine zone” for a configurable period of time, during which whitelisted App Guardians can cancel execution of the message content. If any app can easily add an optimistic component, then why should it be provided by the interoperability solution itself as a second security model?
- If an optimistic approach is implemented into the interoperability protocol itself, it delegates responsibility from the validation layer of cross-chain infrastructure to the application, which should elect “App Guardians” who become participants of the new validation mechanism. This approach performs implicit change from the “shared” to “isolated” security model of the validation layer. The reasoning behind why the isolated security model is not suitable for a validation layer can be found in this L2Beat article and my Twitter thread.
I’m not trying to attack a competitor here, I want to make sure that all trust assumptions and risks are made clear for the Uniswap community, as Celer has been endorsed in the temperature check.
Repeating what I posted in Discord:
Is Uniswap aware/worried about the possibility of a security issue regarding repairing hashes with Beacon Chain?
If this thing is not really a blockchain it means whoever is repairing these hashes is operating the system. We can debate how and to what degree Bitcoin, Ethereum and other blockchains are decentralized. But nobody there is repairing hashes looking backwards whereas here things look decidedly less like Satoshi’s vision.
(I know nothing about “datafinnovation” except they were mentioned in an Arthur Hayes blog post and many of their posts seemed prescient of later issues being revealed.)
Aave, for example, will go on a chain and then remove support if there is security risk (i.e. a reliance on a single bridge, pre-or-post that bridge being compromised). This post suggests serious security risks. I know BSC is attractive considering its TVL, but this could have serious implications, no?
In light of our concerns with Celer, we thought we should propose a solution to the Uniswap Community. To effectively but quickly evaluate cross-chain messaging protocols, we came up with a few key criteria to asses protocols.
- How long has the system been running, and how much value has it been responsible for?
- What degree of security and decentralization do they offer?
- Have their contracts been audited & do they have a bug bounty program?
- Do the contracts have an owner? If so, what can the owner do?
- Does their documentation clearly explain how their system works, not just in concept but, most importantly, in execution?
- What do we know about the team?
It’s a mix of subjective and objective items that have helped us navigate the space. Through this process, Wormhole stood out to us, so we decided to see how the rubber met the road. To best understand the technology, we deployed a copy of Uniswap v3 onto BSC’s testnet and developed a message sender and receiver contract.
Deployed Contract Addresses
- Uniswap Factory Contract: link
- Uniswap Wormhole Message Sender: link
- Uniswap Wormhole Message Receiver: link
- Message Sent to add two additional fee tiers.
- Message Received and added two fee tiers.
The current flow:
- At the Uniswap Message Sender (Goerli) - sendMessage(targets[ ],values[ ],data[ ],messageReceiver). The message is encoded and published. The function is only callable by the contract owner. For testing purposes, it is currently an EOA, but in production, it should be the Uniswap Timelock contract.
- Wait for the guardians to complete their verification process. Since we have hardcoded the
consistencyLevelto 1, it will wait for finality. Once the Verified Action Approval (VAA) is published, then we may proceed.
- At the Message Receiever (BSC testnet) - receiveMessage(bytes[ ]). The contract calls Wormhole’s core bridge contract to verify the VAA. After it verifies the sender is the Uniswap Message Sender contract, it verifies the message has not already been processed and will finally execute the message.
There are three key points. First, the Uniswap Message Receiver must verify the sender is the Uniswap Message Sender. Second, the content received must be the content sent. Third, the message must not have been previously executed.
We would love it if the community reviewed our work. We believe utilizing this configuration for the BSC deployment makes the most sense. If we proceed with this implementation, we’ll conduct an additional test to verify this works with the Uniswap Governance system and coordinate with Ilia.
On a related note, from having spoken with the Wormhole team, it’s our understanding that Wormhole is also preparing a forum post with additional information.
We had the chance to connect with @ilia_0x yesterday over Telegram — we’ve answered the questions you’ve posed to us! The Wormhole community is excited to throw our hat in the ring, and we look forward to engaging with the Uniswap community more broadly
Hi everyone —We’re Wormhole, a leading arbitrary message passing protocol that connects 22 heterogeneous chains. Today, the Wormhole community includes several independent core contributing teams as well as a network of 19 industry leading Validators securing the network. The Wormhole community is investing deeply in a fully trustless future and is working on a Zero-Knowledge (ZK), light client based, backend designed to deprecate the Guardian software. Once enabled, chains can trustlessly validate other connected chains.
Does the bridge support arbitrary message passing?
a) Yes, Wormhole is one of the first and longest serving arbitrary messaging protocols. Since launching on mainnet in August 2021, 185 million messages have been transmitted, with 2 million messages currently generated daily between asset transfers and messaging through organic usage.
Is the bridge secured by a trusted entity, by a multi-sig, or a protocol/set of incentivized nodes?
a) Wormhole is secured by 19 validators (aka: Guardians) who jointly attest to messages. Each message must be attested by at least 13 of the 19 Guardians. Our Guardian set comprises the leading PoS validators, including Staked, Figment, Chorus One, P2P, and more. The complete set of current validators can be found here: https://wormhole.com/network/
b) As mentioned above Wormhole is making significant progress in developing ZK-based light clients to facilitate completely trustless message-passing.
Does the bridge leverage the security of the source chain (e.g. Ethereum L1) or destination chain, or is security provided by another third party entity?
a) Wormhole message security waits for both consensus to be reached on the source chain as well as additional safety features provided by the bridge. Additionally, Guardians run full nodes to protect the protocol against consensus-level exploits in the connected chains and further reduce contagion risk.
Is it possible for a fraudulent message to be passed to the destination chain? If so, are there any recall mechanisms?
a) All messages passing through Wormhole require a minimum of observation and signing by a majority of the Wormhole Guardian set (13 of 19).
b) While there is currently no out-of-the-box recall mechanism for messages, a minority (7 of 19) Guardians may refuse to sign a fraudulent message and thwart an attack.
c) Simple yet customized message recall functions can be built by individual integrators. An integrator would simply build “edge contracts” to introduce a time delay on message acceptance, providing an integrator with an opportunity to recall the message before it becomes effective.
What are the ramifications of fraud to the malicious actor?
a) Wormhole’s Guardians are leading PoS validators and some of the most respected names within the validator community. They collectively represent tens of billions in value staked and carry valuable reputations in the communities where they serve. Should they act maliciously (such as sign or forge fraudulent messages), they risk reputational consequences, external PoS businesses, and ejection from the Wormhole Guardian set.
b) There is little incentive for an individual Guardian to act maliciously. Even if a Guardian were to succeed in forging a fraudulent message, it would not affect the network state because a single signature isn’t enough to establish the super-majority required to gain quorum. Finally, a fraudulent message would be immediately attributable to the offending Guardian to the rest of the Guardian network.
Has the bridge code been audited? By a third party? What attack vectors and vulnerabilities were identified, if any? Have the identified vulnerabilities been remedied?
a) The bridge has been audited 25+ times by leading audit firms, including Certik, Trail of Bits, and OtterSec, and the cohort of auditors continues to grow. You can see the complete list of auditors and publicized findings here. Those 25 audits are in addition to Wormhole’s already rigorous internal auditing standards, where a team of 6 experienced security engineers regularly perform review the protocol’s security.
As these 3rd party audits are completed and issues are sufficiently addressed, we make those audits public.
- January 2022 - Neodyme: Ethereum Contracts
- January 2022 - Neodyme: Solana Contracts
- January 2022 - Neodyme: Terra Contracts
- January 2022 - Neodyme: Guardian
- January 2022 - Neodyme: Solitaire
- July 2022 - Kudelski: Ethereum Contracts
- July 2022 - Kudelski: Solana Contracts
- July 2022 - Kudelski: Terra Contracts
- July 2022 - Kudelski: Guardian
- August 2022 - Kudelski: Algorand Contracts
- September 2022 - OtterSec: NEAR Contracts
- September 2022 - Trail of Bits: Solana Contracts
- September 2022 - Trail of Bits: CosmWasm Contracts
- October 2022 - OtterSec: Aptos Contracts
- October 2022 - Hacken: NEAR Integration
- November 2022 - Zellic: Aptos Integration
- Q4 2022 - Halborn (DRAFT): Wormchain
- Q4 2022 - Halborn (DRAFT): Accounting
- Q4 2022 - Certik (DRAFT): Ethereum Contracts
- Q4 2022 - Certik (DRAFT): Solana Contracts
- Q4 2022 - Certik (DRAFT): Terra Contracts
- Q4 2022 - Certik (DRAFT): Guardian
- Q4 2022 - Certik (DRAFT): Solitaire
- Q4 2022 - Coinspect (DRAFT): Algorand Contracts
- Q4 2022 - Hacken (DRAFT): NEAR Contracts
- Q1 2023 - Trail of Bits (ONGOING): Guardian
We feel Wormhole is well-qualified to support Uniswap’s cross-chain messaging between ETH and the BNB PoS Chain, and we appreciate GFX Labs’s proposed solution. If anyone has additional questions, we would be happy to answer them.
Hey Alex, thank you for your comments. Always great to have discussions with fellow builders in the space. Here are some quick replies:
What do we mean by “Cosmos-consensus Security Model” or “L1-PoS-blockchain security mode”
When we say L1-PoS-blockchain security model, we mean that the consensus algorithm is essentially acting like any other PoS blockchain. The closest in terms of consensus algorithm/mechanism is Cosmos-SDK/Tendermint chains such as Polygon PoS, BNB Chain or Sei Network. This is to stress that Celer uses a time-tested consensus engine and economic security engine, unlike some other multi-sig solutions or solutions with untested economic security models. We are not saying in any way that Celer’s shares the exact same PoS setup with any other blockchain.
Consensus failure != all money stolen
Unfortunately, you are not correct in stating that if the PoS consensus fails (2/3 stake colluding to behave maliciously), all liquidity in Celer’s Bridge can be withdrawn. This is precisely where the model of optimistic rollup-like security model shines. cBridge is built with this model from day 1 and even if the majority stake is acting maliciously, large fund withdrawing (in a cumulative sense) transactions can be canceled by any Celer validator plus a group of app guardians from security firms such as Peckshield and others. So what you described won’t happen. From a more practical point of view, Celer validators are highly reputable validators from the Cosmos community and incidentally share many with deBridge we believe.
Optimistic rollup security: application design pattern or protocol built-ins?
Implementing an optimistic rollup security model can be done in the application domain and in the protocol at the same time. They are not mutually exclusive. For Celer, every single validator can cancel messages that are maliciously passed by the consensus protocol if the message is sent through an OR model. This enhances the security from trust majority to trust-any validators.
In addition, application builders, governance bodies or security firms independent from application builders can run app guardians for specific applications even if they don’t trust any validators in Celer. For example, Certik, Trial of Bits and other security firms that are independent of both Celer and application builders can run app guardians for applications to offer “external security firms certified” services. The architecture is highly flexible with a high level of redundancy.
Hi Wormhole team,
While we are looking at the 2M message per day stats you stated, we noticed that 99.97% of those messages are from a network called Pythnet. If we exclude Pythnet, there are 719 message per day in the last 7 days on Wormhole. To put it into perspective, Celer processes 1958 messages per day during the same period of time.
In addition, we could not find any contract address deployed on Pythnet based on your documentation section describing contracts deployed on supported networks.
With the lack of information and significantly skewed volume, could you please elaborate on the Pythnet use case so everyone in the Uniswap community has a clear understanding on the usage level of Wormhole?
- Who are sending all the 2M messages?
- Where are those messages going? If there are 2M messages initiated from Pythnet per day, does it mean that there are 2M transactions somewhere on a destination chain? We searched from all your public contracts but did not find the matching 2M transactions. Could you please elaborate?
With 2M messages processed per day, one would imagine that there are a lot of unique wallet addresses interacting with Wormhole. Could you please share that data too? This way we can have a more complete comparison.
Hi GFX Labs team,
Just a gentle reminder that we have replied to all of your concerns. Since you have done an evaluation of multiple solutions using the criteria you mentioned and are now championing one single solution, could you please post the evaluation matrix with facts for all the criteria?
With this, different solution providers can clarify if needed, and the Uniswap community and delegates can do a peer review and make informed decisions. Many community members and delegates may reach different conclusions by looking at all the facts.
In addition, we would add the following criteria to the evaluation matrix:
- Are there any closely resembled use cases already live and running in production?
- What is the security track record of the protocol?
- What are emergency response plans or security operation measures in place to prevent and/or limit the scope of unforeseen smart contract bugs and consensus/multisig failures?
However, we feel that we may be having the wrong discussion here. Cross-chain governance is a unique use case and it can be built as a vendor-lock-in-free architecture where multiple cross-chain solutions can co-exist to serve Uniswap with super-charged security properties.
Thanks for the deep analytics and numbers, but I think we as a Uniswap community should evaluate and compare first the security part of different solutions.
I also do not rule out that the following proposals for BNB and other chains may use different cross-chain solutions and not rely on a single point of failure.
I appreciate all the information about the different bridge solutions. I am not familiar with bridges or the security vulnerablilities. I have only heard of the Wormhole bridge from it’s $320 million bridge exploit.
How would an exploit like this impact Uniswap on BSC? Would all LP’s be vulerable?
Is the $320 million that was exploited still an issue that could impact protocols using Wormhole’s bridge solution in the future?
@modong what are the risks related to validator’s using Celer native token as stake to be slashed?
If Celer token prices get low enough, could there be malicious actor’s/validator’s that see more opportunity in the value of the bridge assets vs being slashed in Celer?
Would be good to have a combination of bridges, but due to my unfamilarity in this area I am unsure what is possible.
A Vendor-lock-in-free Universal Governance Mechanism for Uniswap
In this sub-proposal, we present a vendor-lock-in-free Universal Governance Mechanism (UGM) for Uniswap that utilizes multiple cross-chain protocols simultaneously. This architecture significantly enhances the security of Uniswap UGM and, at the same time, eliminates any potential alignment risks down the road.
While we believe that Celer alone can effectively solve the Uniswap UGM use case, as a long-time advocate for open protocols, we propose this solution with the best interest of the Uniswap community in mind.
Which one to choose? Wrong question!
It may seem that Uniswap’s delegates and community must make a choice to select a single cross-chain messaging protocol for Uniswap’s multi-blockchain expansion.
However, if we go down this path, it will be a TRAGIC day for Uniswap community because Uniswap, the most open dex in the world, will be vendor-lock-in to a single cross-chain protocol.
Vendor lock-in carries a multitude of risks, most notably:
- Security risk: Uniswap will be relying on a single cross-chain protocol and if that cross-chain protocol is hacked, Uniswap will also be at risk.
- Interest alignment risk: This lock-in makes it harder for Uniswap to switch out of the solution even if the interests between the cross-chain messaging protocol and Uniswap become misaligned down the road.
But are we asking the right question?
Does Uniswap have to choose one single solution between different ones or can it utilize multiple solutions to achieve a better outcome for the community?
An Open UGM for An Open DEX
Cross-chain governance use cases have the following properties:
- No room for security breaches as governance decisions are mission-critical parameters for the protocol with astronomical financial impacts
- Latency is flexible with the governance process being multi-day if not longer
- Costs are secondary due to the low frequency of governance decisions
With these properties, it is a perfect use case to utilize multiple cross-chain messaging solutions simultaneously.
We propose utilizing multiple cross-chain solutions simultaneously through the use of a “MessageProcessor” contract that acts as the executor for UGM calls to Uniswap contracts on the Binance Chain. This contract would have the following functionalities:
- The ability to receive messages from multiple cross-chain messaging solutions through one interface.
- An internal state machine to track messages received from different protocols and to release the message and trigger governance decisions only when a threshold is reached.
- The capability to remove and add different cross-chain solutions using the same UGM mechanism with potentially different thresholds.
This architecture ensures that even if a single bridge is compromised, Uniswap’s governance process remains intact. Additionally, the Uniswap community can continuously evaluate participating cross-chain protocols with all bargaining power on their side.
We believe this architecture should be implemented to avoid vendor lock-ins, even if the Uniswap community decides to use a single cross-chain solution for now. Celer is committed to a vendor-lock-in-free future and is more than happy to implement this vendor-lock-in-free UGM for Uniswap.