[Chaos Labs] Cap - Risk Briefing

Chaos Labs’ risk assessment evaluates only the suitability of the Cap protocol for receiving stake from Renzo from a mechanism design perspective. We are not qualified to assess the creditworthiness of underlying operators taking loans backed by this stake or the legal agreements protecting Renzo. We strongly recommend consulting specialists in credit risk and legal matters before making any allocation decisions.

As Renzo continues to expand its AVS/Network portfolio, evaluating the unique risk and reward dynamics of each candidate becomes essential to maintaining a robust and resilient restaking strategy. In this context, Chaos Labs has conducted an in-depth assessment of the Covered Agent Protocol, analyzing its economic model, reward structure, and integration of slashing mechanics. This report was prepared in accordance with the AVS Risk Assessment Framework we developed for Renzo. It aims to quantify potential loss scenarios, evaluate key risk dependencies, and assess Cap’s suitability within Renzo’s broader AVS delegation strategy.

Summary

  • Cap is a stablecoin protocol that issues yield-bearing stablecoins through loans backed by restaked collateral. Financial institutions borrow from a vault, while restakers underwrite the loans and earn fees. The system relies on off-chain legal agreements between restakers and operators.
  • A credit event could result in a loss of the delegated amount equal to the entire loan value, which could be significantly higher than the partial slashing typically seen in other networks and AVSs. Restakers must monitor operators closely, potentially needing to add collateral during stress events. Furthermore, the delegated stake remains locked for the entire duration of the loan and cannot be redeployed, which adds duration risk and may impair liquidity for Renzo’s liquid restaking token.
  • Cap economic model introduces a new type of risk to restaking, which depends significantly on off-chain legal agreements between restakers and operators.
  • Chaos Labs strongly recommends obtaining specialized credit-risk assessments to ensure comprehensive due diligence on all underwriting activities before proceeding with any stake allocation. Additionally, we advise we perform a future stake allocation assessment that incorporates the credit assessment results and legal agreement details as key inputs into the recommended allocation mix.

Functionality of the Covered Agent Protocol

The Covered Agent Protocol issues yield-earning stablecoins through covered loans secured by restaking. The protocol is currently live on testnet on Ethereumx featuring cUSD, a USDC-backed stablecoin. Additional stablecoins will be introduced in the future.

Similar to other PMS mechanism, users can mint cUSD by providing USDC at a 1:1 ratio, enforced by a price oracle. Users can then stake their cUSD to receive stcUSD, which generates yield. The reserves backing the stablecoin are held in a vault and deployed as loans to financial institutions such as HFTs, market makers, MEV participants, and others. These institutions, referred to as agents or operators, use the borrowed capital in their own yield-generating strategies. In return, they pay interest on the borrowed capital, which funds the yield earned by holders of stcUSD. To access this capital, an operator must first secure restakers who are willing to underwrite the loan.

The mechanism is similar to traditional lending markets, with the key difference that restakers provide the collateral. Each lending market represents a one-to-one loan between a restaker and an agent. These loans are supported by off-chain legal agreements that are supposed to define the terms, including collateral amount, loan duration, interest payments here called restaker fee, and other conditions. Restakers also set a maximum borrow limit, which caps the amount an agent can borrow against the provided collateral. Cap sets the borrow rate, which is the interest that operators must pay to the platform for borrowing the stablecoin reserves. The interest rate is set using a combination of a minimum base rate, which is closely tied to Aave’s stablecoin deposit rate, and a utilization-based component. This interest translates into the yield of the stablecoin. The protocol also determines the liquidation threshold for each collateral asset type, using a model similar to that of Aave.

Operators use the borrowed capital in yield strategies and, at the end of the loan term, return the principal along with the borrow interest and restaking fee. Any additional yield generated is kept by the operator. The protocol tracks restaking fees, which are then redistributed to the restakers through the restaking platform. The protocol is currently integrated with Symbiotic, with support for EigenLayer coming soon. Interest is paid in the same token as the principal loan. If restakers do not want to wait for loan repayment, they can alternatively claim their fees at any time directly from the collateral pool.

When an operator’s health factor drops below 1, a grace period is triggered, giving the operator a final opportunity to repay the loan. If the debt is not fully repaid within this period, liquidators are allowed to step in, repay the operator’s outstanding debt to the stablecoin vault, and claim a portion of the collateral provided by the restaker. This process initiates a slashing event on the restaking platform. The slashed amount includes both the repaid value and a liquidation bonus, which is transferred to the liquidator. Cap plans to use a separate vault for each restaker within Symbiotic and plans to implement a dedicated operator set per restaker on EigenLayer.

Risk Assessment

Rationale for Our Endorsement

We generally endorse exploring delegations to operators through Cap. Its use case is compelling, and the yield potential is significantly higher than what is currently available across other Networks and AVSs. Several mechanisms stand out as particularly beneficial for restakers:

  • Custom Loan Terms

Restakers have full autonomy to negotiate delegation premiums and maximum borrow rates based on their own risk assessments. This allows for customized risk management strategies.

  • Accredited Operators with Legal Backing

Operators are expected to be sufficiently diligenced by an expert professional credit risk entity. Off-chain legal agreements between restakers and operators provide an added layer of enforceability for loan terms and interest obligations.

  • Stable and Accessible Yield Returns

Restaking fees are paid in the same asset that is borrowed (e.g., USDC or USDT), helping restakers realize stable, predictable returns. Restakers can access their fees on a regular basis. However, in cases where the LTV approaches the liquidation threshold, payouts may be limited.

  • Clear On-Chain Slashing Conditions:

Slashing conditions in Cap are clearly defined and enforced on-chain, reducing the risk of disputes.

  • Option for Active Risk Mitigation

In case of a loan approaching the liquidation threshold, restakers have the option to inject additional collateral to restore the operator’s health factor. This provides a buffer for renegotiating terms or helping the operator meet obligations.

Identified Risks

While we view Cap as a promising protocol with strong yield potential, we also maintain several concerns regarding the nature and distribution of risk. Cap’s design places significant responsibility on restakers, requiring them to perform due diligence, actively manage credit exposure, maintain off-chain legal agreements, and engage in ongoing operator oversight. This level of responsibility goes beyond what is typical in most Networks and AVSs in both risk type and management approach. Primary concerns include:

  • High Maximum Slashing Risk

While the probability of default is considered low, a credit event could result in a loss of the delegated amount equal to the full loan value. This stands in stark contrast to other networks and AVSs, where slashing is typically partial or limited to foregone rewards. From a VaR perspective, this represents a worst-case loss scenario far exceeding the expected slashing modeled in Renzo’s AVS risk framework. Because this risk arises from off-chain credit failure rather than protocol-level misbehavior, traditional mitigations such as StakeSure or slashing correlation assumptions do not apply.

  • Active Position Management Obligations

In the event of a temporary operator default or heightened market volatility, restakers may need to inject additional collateral to restore the operator’s health factor. Although Cap’s guarantor agreement assigns responsibility for maintaining the LTV to the operator, restakers may still be compelled to intervene if the operator fails to act promptly. This creates a need for continuous monitoring, fast decision-making, and maintaining an ETH buffer, demands that may not always be practical given timing constraints, liquidity availability, and market conditions.

  • Heavy Reliance on Off-Chain Agreements

If an operator underperforms or acts maliciously, the only available on-chain recourse is for the restaker to withdraw their delegation. This triggers forced liquidation, which incurs financial losses and is subject to withdrawal lock-up periods, making timely risk mitigation difficult. As a result, the enforcement of loss recovery relies heavily on off-chain legal agreements.

  • Reward Protection Dependent on Legal Agreements

Off-chain legal support is also required to ensure the security of restaker rewards. Several edge cases could impact expected returns, such as early loan repayments triggered by rising borrow rates, depletion of stablecoin reserves due to large-scale redemptions, or a depeg of the borrowed asset. Mitigating these risks depends on carefully structured off-chain agreements and continuous oversight by the restaker.

  • Unclear Governance of Slashing Parameters

For the mechanisms that are on-chain, it remains unclear how key slashing parameters such as the liquidation bonus, grace period, or liquidation window are defined or governed. These settings could be changed in the future without input from restakers, potentially increasing their risk exposure.

  • Restaking Model introduces Liquidity Risk

On Symbiotic, Cap uses a Single Network, Single Operator model. On EigenLayer, it plans to adopt Unique Stake. In both cases, the delegated stake is locked for the full duration of the loan and cannot be redeployed elsewhere. Withdrawing it without loss requires waiting until the loan concludes, plus an additional withdrawal period. This may impact capital efficiency and depending on the size of the allocation may affect the overall liquidity available within the restaking protocol.

Recommendations

When assessing and engaging with operators, restakers should treat the legal agreement as a critical layer of risk management. Cap provides a guarantor agreement template, under which operators are legally obligated to return funds to restakers in the event of a slashing event. We further recommend seeking legal counsel to adapt the terms to the specific risks and operational realities of each operator relationship. Legal experts can help address a range of potential edge cases such as early repayments, defaults, or asset depegs and ensure that appropriate protections are included to safeguard the restaker’s interests.

Given the strong emphasize on the off-chain nature of the relationship between restakers and operators, it is not possible to fully assess risk through mechanism analysis of Cap. Therefore, we also recommend engaging a specialized credit risk manager to evaluate and monitor ongoing risk, and potentially manage bankruptcy proceedings if needed.

Cap introduces a fundamentally different risk profile compared to traditional AVS models. As such, we believe the expected yield should be meaningfully higher to compensate for the elevated risk. Allocations should only be made when the projected returns are commensurate with this new level of exposure.

All allocations should take into account the risks and constraints outlined above, as well as the potential secondary effects on the LRT and its integration across DeFi protocols. These dynamics call for a more thorough assessment to guide informed decisions and shape a resilient allocation strategy. We recommend that Chaos Labs conduct a future allocation assessment to manage liquidity risk, with the stake being illiquid for longer than usual. This assessment should incorporate credit analysis outcomes and the structure of legal agreements as key inputs to inform a robust and risk-aware allocation framework.