Every protocol on VaultMind receives a 1–100 composite risk score built from
six independently weighted dimensions. No black boxes. No opaque models.
Here's exactly how it works.
6
Risk Dimensions
1–100
Composite Score
150+
Protocols Scored
Daily
Score Updates
Overview
The Composite Score
A single number that distills protocol risk across security, liquidity, governance,
oracle reliability, compliance, and operations.
Score Direction
Higher is safer. A score of 85 means the protocol exhibits strong risk controls across
all six dimensions. A score of 30 flags material vulnerabilities requiring institutional caution.
Scores are not rankings against peers — they are absolute assessments against defined criteria.
Score Frequency
Scores update daily using the latest on-chain data, governance activity, and
oracle metrics. Significant events — exploits, governance attacks, large TVL outflows —
trigger immediate score reviews outside the daily cycle.
Weighted Aggregation
Each dimension is scored 0–100 independently, then multiplied by its weight and
summed to produce the composite. Weights reflect the relative contribution of each
risk category to historical DeFi losses and institutional due diligence frameworks.
No Automatic Defaults
Missing data for a sub-signal does not default to zero (unfairly punishing new protocols)
or maximum (falsely inflating scores). Missing signals are interpolated from comparable
protocols in the same category, with a transparency flag in the score breakdown.
Category Weights at a Glance
Smart Contract Security
25%
Liquidity & Market Risk
25%
Governance
15%
Oracle Reliability
15%
Regulatory Compliance
10%
Operational Risk
10%
Risk Dimensions
The Six Categories
Each dimension is assessed independently before being combined into the composite score.
Smart Contract Security
25%
The largest single weight, reflecting that code exploits account for the majority of
historical DeFi losses. We evaluate the full audit history — not just whether an audit
exists, but who conducted it, when, whether findings were remediated, and the gap between
current deployed code and the audited codebase. Active bug bounty programs signal
ongoing security investment and are factored positively. Formal verification, immutable
contracts, and time-locked upgrades all improve the sub-score.
TVL depth, withdrawal velocity, and concentration risk determine whether a protocol
can absorb redemption pressure without cascading failure. We track rolling 7-day and
30-day outflow patterns, the proportion of TVL held by the top 10 addresses, and
whether liquidity pools show structural depth or are propped by temporary incentives.
Stablecoin peg deviation history is also evaluated for relevant protocols.
Governance failure is an underappreciated risk vector. Protocols with single-signer
admin keys or token distributions dominated by one entity can be unilaterally changed
without community consent. We score multisig structure (quorum, signer diversity,
hardware wallet use), token distribution Gini coefficients, historical voter turnout,
and the presence of time-locks on parameter changes. Active governance participation
is a positive signal; silent token concentration is a material red flag.
Multisig structureQuorum requirementsToken Gini coefficientVoter participationTime-lock on changesAdmin key risk
Oracle Reliability
15%
Oracle manipulation is the second most common exploit vector after smart contract bugs.
We assess whether protocols depend on a single oracle or maintain multi-source diversity
(Chainlink, Pyth, Uniswap TWAPs), how frequently price feeds update, and whether the
protocol employs circuit breakers that pause operations during anomalous price conditions.
Historical manipulation events are flagged and permanently weigh down the sub-score.
Oracle diversityFeed update frequencyChainlink usageTWAP protectionsCircuit breakersManipulation history
Regulatory Compliance
10%
Institutional allocators face regulatory obligations that DeFi protocols do not —
yet. We track jurisdictional clarity (where is the team, what entity structure exists),
OFAC sanctions exposure in the protocol's user base, enforcement actions and
regulatory correspondence, and alignment with emerging frameworks like MiCA in the EU.
Protocols operating with clear legal wrappers and proactive regulator engagement
score higher. Anonymous teams with no legal structure are flagged accordingly.
Protocol resilience depends on the team behind it. We evaluate founding team track
record (prior projects, exits, failures), incident response quality (did the team
communicate clearly and act swiftly during past issues?), infrastructure uptime
over trailing 90 days, and team size relative to protocol complexity. Teams that
have navigated incidents well — not teams that have never faced one — demonstrate
the operational maturity institutional allocators require.
Team track recordIncident response historyUptime (90-day)Team size vs complexityCommunication qualityKey-person dependency
Dual-Track Scoring
Two Asset Classes. Two Frameworks.
Payment stablecoins and tokenized institutional products carry fundamentally different risk profiles.
Applying a single framework to both produces meaningless scores — so we don't.
Payment Stablecoins
USDT, USDC, DAI & peers
Pegged currencies where depeg risk, collateral backing quality, and liquidation
mechanics are the primary failure modes. Evaluated using the standard 6-category
framework with smart contract security and liquidity/market risk carrying the
highest weights.
USDTUSDCDAIFRAXcrvUSDLUSD
Tokenized Securities
BUIDL, FOBXX, OUSG & peers
Treasury-backed money market tokens where the underlying fund is 100% short-duration
government securities. Depeg risk and collateral volatility are structurally irrelevant
here — the relevant risks are custodial, legal, operational, and smart-contract
tokenization risks. A separate scoring framework applies.
A Treasury-backed money market token like BlackRock's BUIDL holds 100% short-duration
U.S. government securities in a regulated fund. Scoring it on "depeg risk" or
"collateral volatility" — metrics designed for algorithmic and reserve-backed stablecoins —
produces a misleading result. The actual risks are: who holds the assets in custody,
what happens at redemption, how clean is the regulatory structure, and does the smart
contract layer introduce new failure modes?
Institutional allocators evaluating tokenized T-bills alongside USDT need a framework
that reflects this. A single score conflating both categories either unfairly penalizes
institutional-grade products or obscures genuine stablecoin risks. We built two separate
rubrics to serve allocators with the precision the asset class requires.
Tokenized Securities: Scoring Categories
Five categories replace the standard six. Depeg risk and collateral volatility signals are removed;
custodian, fund structure, and redemption mechanics take priority.
Custodian & Counterparty Risk
30%
Replaces: Depeg Risk / Collateral Volatility
Who holds the underlying assets? Regulated custodians (BNY Mellon, State Street)
with clear bankruptcy-remote structures score highest. Custody concentration,
counterparty credit quality, and fund segregation are evaluated. A tokenized
product is only as safe as the institution behind it.
Is this a registered investment vehicle (40 Act, UCITS, Cayman SPC)?
Which jurisdiction governs the fund, and is there a clear legal opinion on
token holder rights? Unregistered structures with vague legal wrappers
represent material institutional risk regardless of the underlying asset quality.
Fund registration typeGoverning jurisdictionToken holder rightsLegal opinion qualityRegulatory approvals
Redemption Mechanics & Liquidity
20%
Replaces: Liquidity & Market Risk
Can holders redeem at NAV, and on what timeline? T+0 on-chain vs. T+1 traditional
redemption windows matter for institutional cash management. We evaluate minimum
redemption amounts, eligible investor requirements, and whether secondary market
liquidity exists to exit without waiting for fund redemption cycles.
The tokenization layer introduces code risk even when the underlying fund is
pristine. Transfer restrictions, minting/burning controls, admin key risks,
and the audit history of the token contract are evaluated. A sound T-bill fund
running on an unaudited smart contract is still a security risk.
Combines Operational Risk + Oracle Reliability from standard framework
How long has the product been live, and what is the issuer's institutional pedigree?
A tokenized fund from a T1 asset manager with $10B+ AUM and a three-year track
record scores very differently from a first-time issuer six months into operation.
Incident history, NAV reporting frequency, and investor communication quality
round out the assessment.
Issuer AUM & pedigreeProduct track recordNAV reporting frequencyInvestor communication qualityIncident historyThird-party fund admin
Competitive moat: VaultMind is the only platform that distinguishes between
tokenized securities and payment stablecoins in risk scoring. Applying a generic stablecoin
rubric to BlackRock BUIDL misrepresents both the risks and the institutional-grade controls
in place. Our dual-track approach reflects how sophisticated allocators actually evaluate these assets.
Score Interpretation
Grade Bands
Composite scores map to three institutional grade bands. These align with common
risk committee thresholds used by family offices and fund allocators.
A
75 – 100
Institutional Grade
Strong controls across all dimensions. Suitable for inclusion in
institutional portfolios with standard due diligence. Ongoing monitoring
required but no material flags present.
B
55 – 74
Monitored Exposure
Acceptable for limited, monitored exposure. One or more dimensions
show elevated risk requiring active review. Position sizing and
liquidity limits recommended.
C
Below 55
High Risk
Material vulnerabilities present. Institutional exposure not recommended
without bespoke analysis. Often reflects early-stage protocols, recent
exploits, or governance failures.
Dependency Analysis
Composability Risk
DeFi protocols rarely operate in isolation. Dependencies change the risk profile.
How Dependency Chains Affect Scores
A protocol with a strong standalone score of 80 may depend on a lending layer
scoring 45 for its core functionality. In this scenario, the dependency risk
propagates upward. VaultMind models up to three levels of composability depth —
identifying which protocols a given protocol relies on for liquidity, collateral,
price feeds, or settlement.
The composability-adjusted score is displayed alongside the standalone score in
the VaultMind dashboard. For protocols with deep dependency chains, the adjusted
score can differ materially from the standalone figure — a gap that is invisible
in competitors' scoring frameworks.
We also flag "single point of failure" dependencies: cases where more than 40% of
a protocol's functionality would break if one dependency failed. These are
flagged as critical in the protocol risk breakdown.
Example: A yield aggregator routing through three AMMs and a
lending protocol inherits risk from all four. If the lowest-scoring dependency
scores 38, that floor is reflected in the aggregator's composability-adjusted score —
regardless of how well the aggregator's own contracts score.
Data Infrastructure
Data Sources
Scores are derived from public on-chain data, curated off-chain sources, and
continuously refreshed feeds. No proprietary black-box data.