With Fordefi’s integration of Sentora’s Smart Vaults, clients gain access to automated DeFi strategies designed for institutional needs. This overview breaks down how Sentora Smart Vaults work, exploring their core design, execution strategies, and robust risk controls to help institutions deploy capital with confidence.
This is increasingly important because two vaults offering similar yields can differ significantly in their mandates, collateral, venues, leverage limits, and behavior during market shifts. Yield is defined by these operational boundaries, and documenting them ensures these yields are meaningful and comparable across products.
About Smart Vaults
Sentora’s Smart Vaults build on traditional DeFi vaults by adding an active strategy layer. They construct and manage structured positions that involve collateralizing, borrowing, and redeploying capital across multiple venues. Capital flows as follows:
Users deposit an accepted asset into the vault contract.
The vault allocates those assets as collateral or liquidity in an approved protocol.
Depending on the strategy, the vault borrows against that collateral or deploys funds into additional markets.
The position is monitored and periodically rebalanced to maintain target risk parameters such as health factor and leverage ratio.
Unwind logic is defined in advance rather than determined during an event.
The additional layer buys capital efficiency and introduces multi protocol dependency: a requirement for active position monitoring and for clearly defined exit behavior.
Sentora does not hold or control user funds, does not exercise discretion over them, and can only move capital between protocols that have already passed the approval process described below.
The Approved Strategy Set
Four strategy types are approved across the Smart Vault range. Every position a vault holds is an instance of one of them.
Lending
The vault supplies assets into a money market where borrowers draw liquidity against posted collateral, and borrower interest is passed through to suppliers. This is the cleanest way to generate base yield on a highly liquid asset because it requires no directional view or complex position construction.
Returns depend on utilization, borrow demand, the design of the interest rate curve, and any protocol incentive program that is active.
DEX Liquidity Provision
The vault supplies assets to decentralized exchange pools and earns a share of trading fees. Impermanent loss, pool composition, and imbalance dynamics are monitored as primary risk factors rather than as secondary considerations.
Supervised Loans
Supervised lending is designed for blue chip assets that have limited native yield in DeFi, such as WBTC. The strategy uses an asset as collateral to borrow a different, more productive asset, which is then deployed elsewhere. The structure has two legs.
In the primary leg (S1), the vault deposits the blue chip asset as collateral in a lending protocol such as Aave, Morpho, or Euler, and borrows another asset against it, typically ETH or a stablecoin. The collateral may earn a small amount of yield, but the objective of S1 is to create borrowing capacity.
In the secondary leg (S2), the borrowed asset is deployed into a higher yield strategy such as a lending market, a liquidity pool, or another approved DeFi strategy.
The structure is profitable only while the yield generated in S2 exceeds the borrowing cost incurred in S1, which makes the spread between the two legs a monitored variable rather than an assumption.
Leveraged Loops
Leveraged loop strategies use a yield-bearing asset as collateral to borrow a correlated base asset against it and reinvest the borrowed asset back into the same yield source.
For example, the strategy can supply wstETH to a lending protocol, borrow WETH against it, swap the borrowed WETH into additional wstETH, supply that as collateral again, and repeat until the target leverage ratio is reached. The same structure applies to stablecoin yield assets such as sUSDe, syrupUSDC or Pendle PTs looped against USDC or USDT.
The result is larger exposure to the underlying yield while maintaining a position in the same asset. Correlation between the collateral and the borrowed asset is itself a risk control, since pairing uncorrelated assets is a common source of sudden drawdowns.
Sentora's use of leverage is cautious and differs from the leverage typically seen in exchange margin trading. Leverage is strictly managed within defined loan-to-value and health factor boundaries, with conservative leverage levels determined directly by our risk framework.
Risk Infrastructure: From Diligence to Live Controls
Most risk disclosures for vaults focus on initial due diligence conducted before a protocol or asset is approved.
Sentora treats risk management as a continuous control process. Protocol parameters, collateral mixes, oracle routes, and market liquidity can change rapidly, so initial approval is just the first step. This is possible through an end-to-end operational framework, real-time monitoring and strict controls that continuously safeguard active positions against evolving risks. The process involves:
Diligence. A DDQ review of blockchains, protocols, and assets defines the approved universe.
Policy. The risk framework establishes acceptable exposure for Sentora Vaults, Smart Yield clients, and Protocol Vaults.
Controls. Policy is codified into allowlists, LLTV settings, borrow caps, leverage limits, and health factor thresholds.
Monitoring. The Sentora Risk Application tests those controls against live exposure, parameters, oracles, liquidity, and data health.
Escalation. Alerts move into acknowledgement, incident reporting, and Risk Committee review.
Stage One: Diligence Defines What Is Eligible
Before a blockchain, protocol, or asset can support deployment, the Research team prepares an overview covering protocol category, key metrics, capital flow, participant roles, market structure, smart contract architecture, audit history, technical risks, and an initial risk rating. The Risk team converts that into a formal DDQ dossier recording the protocol, token, chain, counterparty context, review scope, decision state, and review timing.
The workflow is evidence-driven and versioned. Dossiers undergo numbered review cycles covering initial assessments, follow-ups, material changes, and re-reviews. Each question maintains a clear status trail, from draft and open to sent, answered, needing follow-up, and closed. This ensures that all questions, answers, supporting evidence, and resulting decisions remain fully auditable after the fact.
Stage Two and Three: Policy Becomes Executable
The framework translates DDQ decisions into controls a system can enforce: allowlists and blocklists, client restrictions, protocol vault directory entries, LLTV settings, maximum borrow caps, collateral tiers, cap utilization thresholds, oracle deviation thresholds, leverage limits, exit maturity limits, and health factor thresholds.
Parameter authority varies by venue: on Aave and Kamino, governance controls core settings (such as caps, liquidation parameters, and oracles), requiring Sentora to monitor settings, evaluate proposals, and engage protocol teams. On Euler and Morpho, curators control these parameters directly, allowing Sentora to establish and govern them. Identical exposures carry distinct control dynamics depending on the venue.
LLTV and maximum borrow cap models sit at the intersection of research, policy, and live risk control. Research designs these models against liquidation buffers, debt capacity, collateral behavior, liquidity depth, and market stress conditions. The Risk team vets the model logic, assumptions, and outputs before a version becomes part of the production control framework.
Stage Four: Controls Tested Against Current Data
Scheduled jobs run daily across position analysis, TVL checks, collateral snapshots, protocol data, client restrictions, and external vault dashboards. Collateral monitoring covers Morpho, Euler, and Kamino, tracking collateral composition, risk tiers, supply caps, and maximum borrow caps, and flagging reserves at the 80 percent, 90 percent, and 100 percent cap utilization bands. This converts a policy limit into a visible risk state before liquidation or exit conditions become disorderly.
Protocol vault monitoring compares current parameters against previous ones and highlights changes in supply caps, borrow caps, oracle fields, LLTV, collateral lists, liquidity, and allocation.
Underneath the monitoring layers sits Risk Radar, Sentora's proprietary risk analytics platform, which computes economic risk signals block by block across more than 40 protocols.
At protocol level, it measures solvency and structural stability through health factor distribution, bad debt, liquidation activity, reserve metrics, and peg deviation.
At vault level, it tracks utilization, collateral composition, and loans at risk of liquidation.
At market level, it monitors DEX liquidity depth, exit cost, net liquidity flows, whale concentration, and volatility distribution.
Stage Five: Detection Becomes Documented Action
Severity levels determine the response path, which is defined before an event:
Severity | Trigger | Response |
|---|---|---|
Info | Normal metrics remain within range. | Logged and visible on the dashboard. |
Warning | A threshold is approached. | Risk team review and increased monitoring. |
Breach | A client restriction or critical threshold is violated. | Immediate alert and a formal incident report within 24 hours. |
Critical | Active threat, liquidation risk, or material control breach. | Risk Committee escalation and corrective action, including position reduction or unwind. |
Where a protocol or asset issuer is involved, the first step is direct contact through established channels to verify the facts before acting. Larger incidents run through PagerDuty into a coordinated response across risk, research, engineering, operations, and client facing teams. Daily reports link current observations to prior observations, so the change in exposure between T-0 and T-1 is recorded rather than reconstructed.
Weekly checkpoints focus on trend analysis, while monthly reviews examine client accounts, mandate compliance, strategy performance, and any variance. The framework itself undergoes a complete review annually or whenever major market shifts occur.
Together, these steps create an end-to-end audit trail, linking the initial diligence behind every decision to the governing policy, enforcing controls, ongoing monitoring, and ultimate resolution. This traceable process is what makes our strategy curation defensible, proving both sound decision-making and operational discipline.
Three Layers of Defense at The Position Level
While the workflow above operates at the platform level, defense in depth applies the same rigors directly to individual positions.
Research and due diligence, before deployment. Every capital allocation requires a formal review. Technical evaluation focuses on audit histories, codebase provenance, and bug bounties. Economic assessment addresses bad debt, de-peg scenarios, revenue sustainability, liquidation mechanics, and oracle dependencies. Counterparty and operational checks examine team backgrounds, multisig configurations, and admin key management to uncover centralization vectors not documented by the protocol itself.
Automated on chain triggers, in real time. On-chain bots execute the control logic where execution has to be immediate. Positions rebalance to maintain target leverage ratios as markets move, which protects the health factor of a position proactively rather than reactively. If on-chain signals such as sharp changes in liquidity, slippage, or utilization breach safety thresholds, the system can unwind a position within the same block. Deleveraging is designed to occur before a health factor reaches a critical level, so that forced liquidation is prevented rather than survived.
Continuous off-chain quantitative monitoring. Active oversight tracks metrics across six risk categories: concentration, liquidity, interest rate, duration, leverage, and correlation. Key metrics include interest rate sensitivity, exit maturity (a stress test estimating the worst-case timeframe to unwind a position), whale concentration, and available pool liquidity to ensure exits occur without significant price impact. On-chain bots handle immediate, single-block execution, while off-chain analytics handle complex evaluations requiring judgment. The Risk Application governs both, connecting automated execution to the approved policy framework to maintain a unified control narrative.
Return of Capital Before Return on Capital
DeFi offers real yield opportunities alongside a set of risks that traditional frameworks are often ill equipped to handle. That gap is closed by treating risk policy as an operating control system that can approve exposure, constrain deployment, monitor current state, identify breaches, and preserve the record of the response.
That is the system described here, and the Sentora Vaults on Fordefi are examples of vaults that utlize this framework.
