Data Reference

Collateral Tiers

Prime, Core, and Edge collateral labels used for registry seeding and display context.

Collateral tiers are registry labels used to seed reviews and explain asset context. They are not the live Asset Composition scoring formula.

The current asset vector scores category-specific dimensions, applies caps, and then applies vault-level portfolio mechanics. See Asset Composition for the scoring path.

How Tiers Are Used Now

UseBehavior
Registry seedingInitial hints for asset review and migration backfills
Display contextHuman-readable labels for familiar collateral quality bands
Analyst review supportStarting point for evidence collection, not a substitute for dimensions
Live scoringDimension scores, caps, overlays, staleness, and portfolio adjustments

Registry labels are kept aligned with the live tier by an automated drift gate in the backend; the handful of deliberate divergences (for example USDT's registry label) are documented exceptions in that gate, each with its reason.

Prime Label

Highest-liquidity assets with long track records and broad adoption. Live tier corresponds to a risk score of 8.0 or higher.

Examples include ETH, WETH, USDC, USDT, WBTC, cbBTC, wstETH, and rETH.

Core Label

Established assets with smaller liquidity, more complex mechanisms, or narrower operating history. Live tier corresponds to a risk score of 5.0 to 7.9.

Examples include DAI, USDS, sDAI, sUSDe, USDe, GHO, FRAX, LUSD, LBTC, OETH, weETH, tBTC, LINK, and ARB. Secondary liquid-staking and restaking tokens (weETH, ezETH, mETH, osETH) sit here on their live scores even where older material described them as Prime.

Edge Label

Unlisted tokens, low-liquidity assets, newer wrappers, assets with explicit Edge overrides, or assets that have not yet passed review.

Unknown assets enter review through conservative fail-safe handling until enough evidence exists for category scoring.

Reading Vault Outputs

Vault outputs may still expose collateral labels for analyst context. Do not compute the asset vector from these labels. Use risk_vectors.asset, asset dimensions, caps_applied, hard_fail_flags, review_status, and risk_score_breakdown where available.

On this page

Raw