v0 previewGrades are draft or provisional: produced under Methodology v0.2.1 from public evidence, pending re-verification and issuer right-of-reply. Nothing here is investment advice, and these are not credit ratings.

Changelog

Every grade change and methodology revision is logged here, appended by a human — monitoring agents flag reviews, they never change grades. Current: Methodology v0.2.1, 23 assets graded.

An unlogged grade change is a bug worth reporting. If a grade on this site differs from the API or from a previous snapshot without an entry below, email dhruv.arunights@gmail.com — errors are corrected publicly.
2026-07-18usdxcoverage queue (candidate, ungraded)→F — tombstoneCommittee decision 2026-07-18, founder-signed via Cowork session. USDX (Stables Labs) entered as a realized-loss tombstone: ~99.3% collapse from its Dec 2024 ATH after a Nov 2025 Balancer V2 pool exploit (~$1M) and a separate crash amid alleged, unconfirmed founder-wallet-drain activity (reported by The Block and AMBCrypto, attributed to on-chain investigators; not confirmed by Stables Labs). The realized_loss_event flag applies, fixing the grade at F. Confidence is medium, not high: the ~99% realized loss of peg is not in dispute, the drain allegation is unresolved. Full evidence trail: ratings/briefs/usdx_tombstone_dossier.md. Deliberate grade-table change — engine/expected_grades.txt regenerated in the same commit.v0.2.1
2026-07-16Methodologymethodology v0.2→methodology v0.2.1Additive regional-expansion amendment (explicit founder instruction): peg_target field with FX-aware peg-deviation model signal for non-USD pegs (defaults to USD, no existing grade affected); documented CEX-breadth + redemption-terms liquidity fallback for thin-DEX assets, auto-capped at medium confidence; new informational regulated_emoney flag (no ceiling) as a positive, evidence-linked Redemption & exit input. No changes to grades, scores, pillar weights, or any existing flag's ceiling. Spec: pivot/09_REGIONAL_EXPANSION_SPEC.md.v0.2.1
2026-07-11Methodologymethodology v0.1→methodology v0.2Added grade-derived collateral haircuts: band base rate multiplied by a confidence multiplier, plus asset-class adjustment and flag add-ons. Every haircut publishes its full derivation string; haircuts are illustrative policy inputs, not margin requirements.v0.2

Machine-readable source: ratings/data/changelog.json in the repository · subscribe via the Atom feed.