Concerns about Custodial Asset Management and Proxy Voting

Let me begin by saying that I believe the Gotchi Vault is a fantastic product with excellent product-market fit, and is a valuable resource for Aavegotchi asset holders. The managers of the Vault are highly valued members of the community who by all accounts act with the best interest of the protocol in mind. Everything said here is in the spirit of protecting the integrity of the Aavegotchi protocol well into the future. I am anticipating that other custodial solutions will emerge that present similar voting power challenges that I will outline here. The Vault just happens to be the main example we have to examine the issue of proxy voting at this time.

The VLT Token and its Role in Aavegotchi DAO Governance

One of the trade-offs to custodying assets in the Vault is that the owners of those assets are entrusting their voting power to the Vault managers. The VLT governance token is issued to the owners of assets held in the Vault during the initial six-month period following the Vault’s launch. This provides a mechanism for these owners to still participate in governance of the Aavegotchi protocol via the Vault managers as their proxy, but is not a 1-1 representation of voting power to assets owned. Should a user remove their assets from the vault, they retain their VLT token and ability to influence the Vault’s vote.

VLT is a decentralized governance token that can be freely traded on DEXs, with a current market value of around $4 USD. Through investing in VLT, anyone could bypass the purchasing of Aavegotchi native assets and just buy VLT if they ever wanted to have a larger influence over the Aavegotchi DAO and base protocol.

The Problems with Proxy Voting

The numbers I’ll outline below suggest to me that the Vault’s voting power creates an outsized influence on the DAO that does not accurately reflect the purposeful will of the owners of the assets contained within. This is in large part because a majority share (perhaps as much as two-thirds) of the assets in the vault are passively owned by parties that do not participate in VLT governance, yet whose voting power is still exercised vis-a-vis the Vault. This being the case, VLT holders who do vote have a disproportionate say over the Aavegotchi DAO relative to the assets they own.

Asset owners are delegating their voting power to the vault in exchange for passive asset management, however there is not an easy way for them to opt to take their voting power back in time to vote on a specific proposal. This is due to the fact that once a DAO proposal is created, a snapshot is taken and voting power is irrevocably allocated to the Vault instead of the asset owner.

The lack of participation by many VLT holders suggests that had these same owners self-custodied their assets rather than placing them in the Vault, they most would likely either be abstaining from DAO votes just as they have abstained from the VLT votes, or are still voting using their non-Vault assets and are unaware of the Vault’s VLT voting process.

The net effect of this is that the Vault’s voting power is transferred from a majority share of passive asset owners to a minority of VLT holders who actively participate in governance. As a result, it takes a very small number of VLT holders (sometimes as few as 2 or 3 wallets) to sway the majority of the VLT vote and in turn wield a near-majority of the Aavegotchi DAO voting power.

Results of Recent DAO and VLT Votes

Here is a summary of VLT and GHST voting data from the most recent coreprops as well as two recent sigprops where voting was substantially closer than the coreprops:


Key Observations and Thoughts

  • The average total GHST vote for the coreprops was about 17.2 million GHST worth of voting power (sigprops are 13-14 MM GHST) The Vault currently holds a total of 6.8 million GHST worth of that voting power - split nearly evenly between two voting addresses - representing a roughly 40% share of the overall GHST vote for the coreprops (roughly a 50% share for the recent sigprops).

  • In one recent instance where the VLT vote was relatively close, the Vault Managers split the Vault’s voting power between its two voting addresses. This is a useful approach for close votes, however it has limitations in my view, as it is not feasible for the Vault to split its vote in a precise manner according to the VLT vote. It’s either 100%, ~50/50 split, or abstain.

  • The average VLT voting power that was exercised on recent votes was ~58,600 VLT. The top 3 VLT holders collectively hold over 30,000 VLT- representing the majority needed to determine the outcome of those votes.

  • There are currently over 2,000 holders of the VLT token. An average of 88 voters participated in the VLT vote in the recent batch of coreprops- representing about ⅓ of the total VLT voting power. The Vault does not have a quorum requirement for VLT votes on Aavegotchi DAO proposals.

A Possible Solution

The spirit on which the DAO operates is that the asset holders express their purposeful will by actively participating in voting. Rather than having custodians such as the Vault vote on behalf of the owners of the assets it custodies, is there a way to allow DAO participants to retain their voting power and cast their votes directly and individually even while their assets are held within the Vault?

As part of such a solution, the DAO would need to require any custodial solutions to disclose their wallets to the DAO, and those wallets would be excluded from DAO voting to prevent double-voting. At the same time, the DAO would instead credit the asset owner’s wallet with the appropriate voting power for their assets held by a custodian.

In short: everyone votes normally, regardless of whether your assets are in the vault or out of the vault. The vault does not vote for you. Everyone has exactly the amount of voting power the DAO allows based solely on the Aavegotchi assets they own and nothing else.

This does require the asset owner themselves to participate in DAO voting in order to obtain the XP rewards, rather than passively earning XP via proxy voting. I see this as a positive side effect for the protocol. (Correction: Vault gotchis do not receive passive XP when the Vault votes)

Some integration with the Vault’s contracts and the Aavegotchi DAO would obviously be needed for snapshot to calculate the correct voting power. I have not investigated whether such a solution is technically feasible.

Final Thoughts

Proxy voting poses a risk to the credibility of the Aavegotchi DAO voting process in that the integrity of the DAO vote becomes dependent on trusting an outside party. In the example of the Vault, the VLT token is acting as a de-facto governance token for the Aavegotchi protocol, in addition to its intended use case as a governance token for the Gotchi Vault protocol.

At best, this is a bad look for the Aavegotchi protocol and presents an obstacle to achieving truly on-chain governance in the future. In a worst case scenario, a proxy voting system with substantial voting power could leave the Aavegotchi DAO and its $7 million+ USD treasury vulnerable to a Sybil attack by an adversarial party.

I’ve tried to be as thorough and unbiased in my thinking on this as I could here, so at this point I would leave it to the community to discuss constructively as to whether the concerns I’ve raised here are warranted and what, if any, proposals should be introduced by the DAO.

The Short Version (TLDR)

  • Proxy voting implemented by the Gotchi Vault via the VLT governance token is not an accurate representation of the purposeful will of the owners of the Aavegotchi assets contained in the Vault. This is due to low participation of VLT voting and it not being a 1-1 representation of Aavegtochi asset owners who are utilizing the vault for its amazing asset management services that it provides the community.

  • The voting power held by the Gotchi Vault is significant - near a majority of the overall DAO voting power and can be influenced by 2 -3 VLT holders. While we may trust these people to have integrity and Aavegotchi’s best interest at heart, we do not know for certainty who else may hold a large share of VLT tokens in the future. This poses a great risk to the DAO as their VP has the power to sway any vote in their favor.

  • A simple solution would be to disallow proxy voting in the protocol and to require all voting power to be allocated to the asset owners instead of the custodians managing those assets. This would require some technical integration between the DAO and the custodians.


I feel that the vault is a valuable enough community resource, that we should pay for the dev work to accurately represent votes within the vault, and to do the same for any other service that comes along and needs help to integrate properly with our voting system.

As the game evolves, and especially in alpha and beta, the voting is going to keep getting messed up some way or another, and as we want no delays, it would be a smooth move to have someone on retainer to fix it for us anytime it gets out of whack or someone makes a cool new service that somehow confuses the asset ownership query.


I don’t wanna dive too much into the subject because it’s too nuanced for me to give it the appropriate time, but I agree with you in that proxy voting makes no sense if the voters don’t actively make opinions known first.

If you don’t vote internally it shouldn’t count as a vote externally, IMO. Can a proxy vote system not have a second and third wallet and divide voting power up based on the internal snapshot?

That way people can vote they way they feel instead of just a big block of votes going one way even if the vote is split 60/40 internally.

I hesitate to respond to things like this because of my lack of dev knowledge but the problems with the system the way it is are glaring and need addressing before we advance past the point of no return.


I think you have a good point. It will strongly encrease the envolvement of players and Frens using the vault. That would be a good thing imo because it will create a stronger community.


Excellent write up ser Mikey!

I couldn’t agree more with pretty much errything you said and the concerns outlined are more than genuine and should be looked into fixing before things are out of our hands (if they aren’t already).

These same concerns have been addressed numerous times in the community before but maybe with the wrong undertone. That’s why it is pleasant to see such a detailed and diplomatic way of addressing it. For that, I salute you fren!

Lastly, it should be a no-brainer (and I don’t get why it isn’t) that:
No Voting = No VP
No Active Participation = No XP

My technical insight is not good enough to dive deeper into the mechanics, but yes, your (our) concerns are more than warranted.


I agree that we shouldn’t be giving XP to those who don’t actively vote, especially when rarity farming rewards are allocated to the XP leaderboard.

I support the proposed fix of disallowing proxy voting and requiring all voting power be allocated to the asset owners.


Is there any scenario where support of voting power delegation is allowed or accepted? I personally participate in the VLT internal votes as well as the AavegotchiDAO votes to account for my split VP. But I don’t think individuals who see the value in owning Aavegotchi assets but don’t have the free time or inclination to be involved in governance should have their VP nullified if they are comfortable with the entity they are delegating in their stead. Do we want VP to be allocated to one human = one vote or do we support VP allocated based on asset ownership? If a whale wanted to enter Aavegotchi but is told they have no representation or voting rights unless through direct interaction, I wonder how that would be interpreted? If they are willing to pay for asset entry, their VP should be theirs to allocate as they see fit, direct or delegation.

Why do we think the Vault has been so successful in asset/VP aggregation? The VLT token doesn’t have deep liquidity available, making it difficult to gather large voting blocks. For this same reason, there wasn’t a direct cash incentive to deposit assets and earn the VLT bribe style as it as hard to liquidate if desired.

Is this an external opinion on the Vaults operating mechanics? Otherwise, if internal users are feeling slighted or misrepresented, don’t the Vault users have the right to withdraw their assets and collapse the Vault’s VP aggregation, regardless of the VLT distribution? Why isn’t this discussion happening among the current Vault users?

It’s quite late so I hope I’m not coming through ignorant, though maybe exhausted. I just have strong feelings about the benefit the Vault has contributed to our entire ecosystem and continue to only see the downsides discussed. Who else has been able to bend the will of QiDao and Dystopia on behalf of all $GHST holders? I believe the positives far outweigh the negatives and we are simply waiting for more ecosystem competition to develop and offer dilution opportunities for the Vaults current aggregation. If they lost the bulk of their asset custody tomorrow and their VP collapsed, would anyone still care about the VLT distribution or how active their internal voting community is engaging?

Thanks for the great writeup and reference data. I look forward to seeing more discussion on this topic. :heart:


I’d frame this discussion on the effectiveness of the XP incentives rather than proxy voting itself - which I also don’t have much of a problem with.

Proxy voting is useful for people who don’t want to directly vote for a variety reasons. Be it not enough time, engagement level or gas fee expenses (not the case here).

However, the original intent of the XP incentives was to reward voters for putting in the time and effort to understand proposals and vote accordingly. It seems a bit unfair if a passive proxy gets the same incentive as an engaged, individual vote.


What an excellent question! My impression of Vault voting is purely external. The Vault’s process are their own and if people don’t like it they can leave.

But if people don’t have time nor the inclination to vote, they shouldn’t get xp.

If people want to give away their VP they should be able too, but giving XP for this doesn’t really seem to be aligned with the spirit of XP…

Do we have two separate issues here? XP distribution and DAO Voting Vulnerabilities? Should the Vault’s internal voting problems also be considered separate from concerns over DAO voting procedures?

Aavegotchi is so complex… this is obviously an investor’s game, so we should prioritize asset ownership when voting. This is obviously a socializing/governance/MMO game, so we should prioritize individuals when voting. Should the DAO differentiate between proposals that affect social or financial aspects of the game, and use specific voting methods accordingly (Aavegotchi asset based or PoE NFT snapshots)?

Perhaps the first (and most appropriate) step is to put pressure on the Vault and create incentives for them to be more supportive of DAO voting integrity…


On a core level, the XP thing is where we actually go against our intent with this situation. It really is THREE issues.

  1. Is proxy voting ok, and do we care if it is delivered in the correct ratio? ← should figure that out and pick a nuanced stance and stick with it. It should be informed by the answers to 2 and 3.

  2. Are we concerned about the ability to own 51% of the vote? ← Extremely dangerous situation in a trustless environment. Look at the DAO facet and tell me you aren’t worried about what could happen if a bad actor had 51% for even a moment.

  3. We give XP for voting and not voting is getting XP. <–this part is going directly against something we have a stated intention for, and is essentially botting or exploiting the system, to get the XP without the participation. If anything gets resolved here, besides some interesting discussion on how to use technology, fixing this should be doable. We could simply not give XP to the owners of gotchis that didn’t vote in the vault vote. This is doable, and is a 1st step towards simply getting the accurate counts as well.

Just in case people haven’t read the wiki on the DAO… we are currently in Cocoon phase -

And metamorphoses is next. As you can see in the next phase, we better have questions one and two sorted or we are exposed. UI 2.0 is where that becomes a possibility. The in DAAP voting is next, and in Realm voting is the final stage.

This topic is not really about the vault… it is about setting intents as a community for how we handle proxies, how we protect against 51% attacks on our DAO facet, and how we hand out XP for voting.

The vault just pushed technology to a point where now we have to think about these things, and frankly we should have handled this immediately, because if there is one, there will be more.

Here is the first example of what is coming, that will cause havok with voting…

Escrow OTC trading - items in escrow will be owned by a 3rd party temporarily. We currently have a 3rd party offering this service and they will end up with a lot of votes if we do not get ahead of this.


I have a question, would lazy people always vote for the first option just to get XP?

Proposals should set the first option to abstain


I think that lazy voters will just vote for whatever option is currently winning the vote, and in a way that’s fine… Just getting XP is the reason we all vote in the first place, take that incentive away and I don’t know if we would still be the most active DAO in the world tbh.

1 Like

i haven’t read your whole topic in depth. but i know what you are referring to and i have thought about this subject a lot. i agree with the idea to not have proxy voting and actually assign all the voting power back to the true owners :+1:


Hey there,

Thanks for bringing up this topic @MikeyJay.

And I do absolutely agree that the actual situation is not in favour of the active voters and participants of the DAO.

@HARDKOR already pointed it out, the main question is if proxy voting is okay in general. I don’t think so!

Any investors in Aavegotchi should be aware that XP is a reward for ACTIVE participation within the DAO. Your responsibility to vote about the further development of the ecosystem should not be delegated to third parties, that maybe reflect your interests.

The problems arising with the growth of the third party and its voting power are threats that need to be addressed now. We, as the community, should not be in a situation to ask ourselves “how is this third party going to vote and decide the outcome of the core prop”. That is what we are in for right now. Vault has got so much voting power that props are decided by their vote/s. That is too much power for one party to accept.

I would like to see if we are already at a point of no return, as if a prop is going to be set up with the intention to not allow proxy voting, the Vault most probably will decide about the outcome.

That being said, the Vault provides a service for the ecosystem that is definitely demanded as we can see in the growth of the Vault.



With the “on-chain voting directly calling smart contracts” future in mind, then yes allowing a 51% proxy vote is crazy.

In Tradfi, >30% of total available voting power is roughly where governance starts getting weird. A lot of jurisdictions mandate you just put in an acquisition offer at that stage :).

It would be good to start landing the principles to move into the Metamorphosis.

If the custodial entity doesn’t (or has no need) for an equivalent DAO governance voting process, voting power should be counted back to the individual owner. A bit like GHST in Quickswap LPs.

A good custodial governance process will reflect voting as per the vote split delivered within the custodial solution. Even then, if a custodial entity has >30% of total voting power, you’d have to consider limiting proxy voting to only “smol or medium” proposals in Metamorphosis.


Great to see this discussion happening, tho not sure why everyone keeps saying that people are getting passive XP from the vault voting. This has never been the case; Pixelcraft removes the vault addresses from the XP distribution and only drops XP to user’s vault gotchis if they voted on the actual Aavegotchi snapshot proposals themselves.


The same is true for attending events. Only individuals who fill out the typeform themselves receive XP. Vault gotchis do not receive passive XP.

1 Like

This is good to hear. I think it may still say you get XP in some of the vault documentation?

I was looking into some other resources and stumbled across this… <-I wonder if this could be used to outsource our voting maintenance. Have a contract to actively make sure that everything works as intended. Every time a change is made or a service is launched, they audit it and tell us if anything changed. This would stop us from not finding out until we make a proposition. They would alert us to the issue and send us a best effort on a potential fix. This would keep things on the up and up and avoid things needing to get really bad before we are OK with bothering PC about it.

It might also be useful for our endless info needs. If we are having trouble sourcing the info from the community, or we need an impartial answer, decentralized analysts is just the thing, IMHO.

They can be paid in GHST, too.


Thanks for bringing this up. I was not aware of this and it sounds like many others were not as well. I added a correction to the original post to avoid further confusion on this.


I am not in favour of delegated voting in a situation at this time .
If there were many other organisations other than just the vault competing for players votes I would feel ok about delegation but clearly the amount of voting power put in the control of so few hands is frightening .
On the subject of reward for participation petting remains a sore point which I am surprised has not been raised in this thread .
The vault as well as other petting services have removed the need for Gotchi owners to participate and has there for incentivised lazy behaviour and convenience over active participation .
I have always considered petting to be a key component of the game, which is now playing out with channeling results being determined by Kinship.
The ideal situation in my mind is that the Gotchi owner who set their alarm clock or by what ever means made the effort to actively engage with the platform to pet are rewarded for engagement .
Otherwise remove petting, under the current model it is not achieving the desired result of engagement .