The 12 Dimensions
Each dimension captures an independent axis of protocol risk. Weights reflect empirical correlation with exploit outcomes across 200+ historical DeFi incidents. 11 of 12 dimensions are fully open — formulas, weights, and data sources are published. D7 (Adversarial Resilience) uses proprietary methodology.
Active Dimensions
Permission models, admin surface, reentrancy protection, and authorization boundaries. #1 exploit vector by dollar loss in DeFi history.
View full detailsRobustness of economic model against flash loans, MEV, liquidation cascades, and incentive misalignment under stress.
View full detailsOracle architecture, manipulation resistance, staleness protection, fallback mechanisms, and feed redundancy.
View full detailsActuarial credibility model: Z-weighted deployment history, TVL-days, audit depth, bounty coverage, and incident history.
View full detailsAdmin key concentration, timelock coverage, upgrade mechanisms, and Maximum Extractable Value by Admin (MEVA).
View full detailsMeasures protocol resilience under independent adversarial security research. Higher scores reflect stronger resistance to sophisticated attack methodologies.
View full detailsIncident response speed, deployment hygiene, key management, monitoring infrastructure, and emergency history.
View full detailsProtocol-intrinsic composition: external dependencies, cross-contract interactions, and integration surface area.
View full detailsExtrinsic systemic risk: upstream dependency failures that propagate into this protocol, and downstream blast radius if this protocol fails.
View full detailsCompiler version CVEs, library dependencies, build reproducibility, and proxy pattern risk.
View full detailsConditional Dimensions
These dimensions activate only when relevant to a protocol's architecture. When active, their weight is redistributed from the base allocation.
Conditional: liquidation cascades, holder concentration (Gini), exit queue analysis. Active for protocols with tradeable positions.
View full detailsConditional: bridge trust model, sequencer risk, fraud proof maturity, cross-chain replay risk.
View full detailsWeight Distribution
Active weights sum to 100%