Draft: Treasury allocation to Castle Vault

Hello Mango DAO!

TLDR: We propose depositing 750,000 USDC of the Mango Treasury to the Castle Vault, a new DeFi savings account built specifically for DAOs.


  • The Castle Vault safely sources yield from the most trusted lending projects in Solana (currently Solend, Port, and Jet – Mango coming soon).
  • It prioritizes minimizing risk over maximizing returns. Funds are only sent to open-source, audited protocols and multiple risk management measures are implemented.
  • It is already integrated with Realms – no dev work is required from the Mango team.
  • There is instant liquidity upon any redemption requests (i.e. no lockup period).
  • We’d love for this to be the start of a long-term partnership between Mango and Castle. Castle will integrate Mango soon to reinforce Mango’s lending liquidity. Castle will waive any future fees on Mango’s deposits into Vault that are received prior to June 30th, 2022.


We are the team from Castle Finance and we would like to propose deploying Mango treasury funds to our newly launched DeFi savings account, Vault. This is a follow up to Brian Smith’s “Mango Markets: DAO Treasury Allocation Strategy”, where he suggested a 5% allocation to Castle’s Vault (mislabeled as “Temple Finance”). We propose an initial allocation of 1%, or 750,000 USDC. This can be increased to the target allocation over time as trust is gained in the Castle protocol and team.

We are grateful for the support we have received from members of the Mango community leading up to our mainnet launch, namely Brian. A partnership between Castle and Mango embodies the ethos of the ecosystem. The Mango community will be able to leverage Vault to safely earn yield on stablecoins during a turbulent market, while Castle will benefit from the reach that the support from Mango would garner.

Because Mango is one of the only truly decentralized DAOs on Solana, and a model customer for Castle, we’d like to offer a fee waiver on any funds deposited before June 30, 2022.

About Castle Finance

Castle is building the next generation of crypto-native products that enable organizations to deploy their assets in a risk-conscious manner. We empower protocols, companies, and DAOs with the financial tools they need to function at their highest capacity and execute on their visions. The Castle ecosystem of products addresses the full array of treasury assets and core financial functions. We work closely with prominent DAOs to create tailored solutions, simplifying treasury management and freeing up bandwidth so web3 builders can focus on development and growth.

About Vault

When stablecoins are deposited into Vault, they are forwarded to a diversified set of lending markets. Vault automates active management of stablecoin lending, rebalancing between sources to optimize the yield earned. DAOs utilize Vault for various goals, from funding operating expenditures to growing corporate savings accounts and protocol insurance funds.

Vault is built on top of the most reputable lending primitives in the Solana ecosystem. We source yield from the following protocols:

  • Solend
  • Port
  • Jet

Additional integrations are currently under development (including Mango Markets!)

Emphasis on Security, Risk Management & Liquidity

In order to minimize the likelihood of principal loss and always ensure liquidity, Vault implements the following risk management measures:

  • Overexposure Control: Vault limits exposure to any single underlying protocol (risk-weighted concentration limits).
  • Underlying Liquidity: Vault ensures that there is liquidity at all times by never lending funds to money markets near a 100% utilization rate.
  • Protocol Security: Underlying protocols are open-source, audited, have verifiable builds, and have decentralized deployment keys.
  • Audit: Bramah Systems, an expert in distributed ledger security, has conducted a full audit of our protocol’s codebase and found no unresolved issues. As we develop additional features, future builds will be reviewed in the same manner.
  • Emergency Brake: Castle’s team will be able to “emergency brake”, which immediately withdraws all funds from underlying protocols into Vault’s reserves.


A deployment from Vault can be made immediately as we are currently live on mainnet. We suggest starting with a 750,000 USDC pilot and periodically reviewing performance. The Castle team is currently working on a comprehensive analytics dashboard that will allow the Mango community the ability to monitor the performance of their allocations with ease.

Additional Resources

Please check out our blog where we post insights and research: blog.castle.finance

More information can be found on the Castle Finance website and documentation.

Why Castle and Vault

At Castle, we are building DeFi products with a strong emphasis on risk management in order to minimize any chance of principal loss. We prioritize prudent management over possible returns. This starts with identifying DeFi’s unique risks, implementing mitigation strategies, and THEN optimizing yield earned within our risk-management frameworks. We are collaborating with ecosystem participants to research systemic risk, define best practices, and create innovative solutions to DeFi’s most common problems.

The Castle team consists of individuals who are passionate about supporting projects building an open, fair, and decentralized future. We are assembling a diverse team with backgrounds from machine learning, software product development, and traditional finance and research.

In short, we believe that Vault is an ideal product for earning stablecoin yield for the following reasons:

  • Open Source - We are completely open source, and only build on top of audited and open source protocols with strong track records
  • Instant liquidity - users can always withdraw their funds with no lock up
  • Security First - we like to think of ourselves as the most risk-conscious team in Solana DeFi. Along with being successfully audited by Bramah systems, members of the Castle team possess the expertise to identify security flaws in code, bringing potential bugs to the attention of several teams in the Solana ecosystem
  • Risk Managed - we use the principles of diversification and overcollateralization, prioritizing principal protection before absolute returns
  • Automated Active Management - within our risk management constraints, we periodically optimize and rebalance across our integrations to ensure you’re earning the maximal, safest yield. This frees up bandwidth for communities to focus on what they do best
  • Realms Integration - Mango will be able to manage allocations through their multisig wallet on the Realms/SPL Governance interface that they have used for over 65 community proposals to date.

Thank you for the consideration! Would love to field and answer any questions here or on Discord.



I am in favor of a small pilot allocation. The Castle team has been doing great work in Solana’s treasury management space. It is important to help support projects that share Mango’s values. Castle’s open-source ethos (including only-OS integrations) and their SPL governance interface are helping advance the Solana ecosystem in the right ways.

Mango will directly benefit once Castle integrates our platform. This will increase liquidity for relevant markets so traders can get the best rates on margined assets. I personally am sick of the elevated borrow rates on Mango’s SOL and can’t wait for Castle to help bring that down.

I understand the community’s concerns around security but Castle seems like the safest option: only audited and open source projects with concentration limits. Realistically, the only way to demonstrate security is a track record. A small allocation from Mango will help prove Castle’s safety and pave the way for a more meaningful allocation (and yield) for Mango down the road.

The Treasury will be holding >$50mm of USDC in reserves for the next several years. It would be great to earn interest on that. Starting small with a $500k allocation to Castle will be an important pilot test. We can reassess in 6 months based on market yields and any security concerns.


We appreciate the support Brian!

As an update for everyone here on the timeline for this proposal:

  • 6/17-23: Draft open for questions and comments here or in discord
  • On the 23rd, I will join the dev call to answer any other questions sync
  • On the 24th, we’ll finalize the proposal with all the feedback and submit it for voting

As always, happy to address any questions or concerns!

Hi everyone! Just popping in here again to give an update on how we handled the recent Solend situation:

This incident shows not only our differentiation from other yield aggregators (which were full risk-on in solend), but also compared to allocating to lending protocols directly (where it would have been difficult to avoid the risk)

Let me know if you have any questions–I’ll also be in the dev call today!

I’ve just submitted the proposal for voting: Realms

Would like to request that if you are inclined vote no, please DM me or post in Discord what your concern is :slight_smile:

1 Like

It would be great to see an integration to Mango Markets itself, I see that you mention it is coming @charlieyou - do you have an indicative timeline?

At GM we don’t think this necessarily is worth the security risk (as Mango core devs haven’t audited code), and currently parking the $500k USDC on Mango would yield more than parking it within Castle. Also noting that $500k would represent 50% of your current TVL, of course there aren’t that many DAO’s with the same size treasury as Mango, but it’s quite a large %.

Could you elaborate on this mechanism: * Overexposure Control: Vault limits exposure to any single underlying protocol (risk-weighted concentration limits).?

1 Like

Hey Klossie,

We expect to start the integration next week, and get a version on devnet by end of July. We’ll get it audited and hopefully be able to ship to mainnet ~2w after that.

As for an audit by Mango, we didn’t want to put undue burden on the dev team, and instead got an audit from Bramah (who had previously also audited Saber, UXD, Set, dYdX, and others). But if this is viewed as a pre-requisite, we’re OK with delaying the vote until that is able to happen. All the code is open source here: GitHub - castle-finance/castle-vault at main

Re. over-exposure controls: there is a hard-limit on the max percentage that can be allocated to an underlying, currently set at 60% for all protocols. We’re working on a quantitative methodology to assess risk on a protocol-by-protocol basis that will inform this setting and others.

I think a small pilot program with a team that shares our values is fine.

But I honestly don’t think any DeFi protocol right now has good risk/reward for USDC deposits. Solend just had a big scare with a massive borrower and we got pretty close to getting a socialized loss scenario on USDC depositors. Mango has fairly low yield and suffers from a lot of the problems Solend has (at least in v3). So, from a risk/reward perspective I don’t think this is great. Rates above 10% are probably more reasonable to start looking at given the risks of current protocols.


Had a quick look at the program deployment @charlieyou . The anchor build doesn’t seem to be verified and your Upgrade Authority looks like it’s some kind of hotwallet.

All in favor of a pilot in general, but would wait until you have those two issues sorted out.

1 Like

Hey Max,

We’re on an old version of anchor, and thus unable to use the verification. We’re upgrading right now and then seeking another audit, so this will change soon.

And the upgrade authority has been switched back to a Ledger, that was indeed an oversight and won’t happen again.