Arunights Research — Collateral Review Evidence Pack

Generated 2026-07-19T03:07:48+00:00 · Methodology v0.2.1 · Ratings as of 2026-07-11 · Portfolio: Demo book (synthetic)

Summary

Gross book value$27,950,000
Risk-adjusted collateral value$24,440,000
Haircut drag−$3,510,000 (12.56%)
Exposure-weighted score78.5

Policy checks

No realized-failure (F) exposureFAIL
Single instrument <= 25% of grossFAIL
C+ and below <= 10% of grossPASS
Low-confidence grades <= 15% of grossPASS

Positions

InstrumentAmount (USD)Grade HaircutAdjusted (USD)
BUIDL
BlackRock / Securitize · RWA wrapper · conf high · draft
10,000,000A6.0%9,400,000
Derivation: base 4% (A) x 1.00 (high confidence) + 2pp class + 0pp flags
Falsifier: Any gap between token transfers and fund-share registry reconciliation would cut Onchain Implementation.
Evidence: [1] [2]
USDC
Circle · Yield dollar · conf high · draft
6,000,000A4.0%5,760,000
Derivation: base 4% (A) x 1.00 (high confidence) + 0pp class + 0pp flags
Falsifier: Evidence of reserve composition drifting from cash/T-bills, or loss of a primary banking/custody relationship without replacement, would cut Backing and Counterparty scores.
Evidence: [1] [2] [3]
USDY
Ondo Finance · Yield dollar · conf medium · draft
3,000,000B+12.5%2,625,000
Derivation: base 10% (B+) x 1.25 (medium confidence) + 0pp class + 0pp flags
Falsifier: Verified instant-redemption facility at scale would raise Redemption; any commingling finding would collapse the grade.
Evidence: [1] [2]
USDe
Ethena · Yield dollar · conf high · draft
4,000,000B17.0%3,320,000
Derivation: base 14% (B) x 1.00 (high confidence) + 0pp class + 3pp flags · Flags: No retail redemption (institutional mint/redeem only)
Falsifier: A major hedging-venue failure handled without holder loss would raise Counterparty; a prolonged deep-negative funding regime draining the reserve fund would cut Backing materially.
Evidence: [1] [2] [3]
sUSDe
Ethena · Yield dollar · conf high · draft
1,500,000B-23.0%1,155,000
Derivation: base 20% (B-) x 1.00 (high confidence) + 0pp class + 3pp flags · Flags: No retail redemption (institutional mint/redeem only)
Falsifier: Cooldown removal or demonstrated orderly mass-unstake under stress would raise Redemption.
Evidence: [1] [2]
syrupUSDC
Maple Finance · Yield dollar · conf medium · draft
2,000,000B-25.0%1,500,000
Derivation: base 20% (B-) x 1.25 (medium confidence) + 0pp class + 0pp flags
Falsifier: A full credit cycle with sub-1% realized losses would raise Backing; borrower-book opacity increasing (or withdrawal queue growth) would cut it.
Evidence: [1] [2]
mBASIS
Midas · Yield dollar · conf medium · provisional
800,000C+37.5%500,000
Derivation: base 30% (C+) x 1.25 (medium confidence) + 0pp class + 0pp flags
Falsifier: Published third-party verification of strategy positions and drawdown controls would raise Backing and Counterparty.
Evidence: [1] [2]
lvlUSD
Level · Yield dollar · conf low · provisional
400,000C+55.0%180,000
Derivation: base 30% (C+) x 1.50 (low confidence) + 0pp class + 10pp flags · Flags: Reserves not fully verifiable
Falsifier: Published, verifiable breakdown of reserve deployment with conservative venue caps would raise Backing and confidence.
Evidence: [1]
xUSD
Stream Finance · Yield dollar · conf high · tombstone
250,000F100.0%0
Derivation: realized loss event -> 100% · Flags: Realized loss event — tombstone, Undisclosed external manager, Hardcoded-par oracle while used as collateral
Falsifier: None. Realized loss event.
Evidence: [1] [2]

Integrity

Every figure derives mechanically from the two hashed inputs below — re-run engine/audit_pack.py against them to reproduce every derived figure and the manifest hash (generated_utc, the run timestamp, is the only field that varies; it is excluded from the hash — preimage rule stated in the manifest's hash_scope field).

ratings.json sha256: e50b5dbfc3337a5d712d974802bee8389672889f5c45584251c938003d3c94a3
portfolio.json sha256: ecda53a92879db16be24fc58eec337f3281dc8758e21a79a8231cd3d7e2d34fb
manifest sha256: 2a4268ba977b617a3f4c2e75097a35a7d64aa917016f3ba9347ac0a7b30db203

Disclaimer

Research opinions under a published methodology. Not credit ratings, not investment advice, not margin requirements. Grades marked draft/provisional are pending verification and right of reply.