Blockchain registry — public-record flags, not a legitimacy score

Built 2026-08-22 from research/blockchain-registry.json. Overlay; excluded from the financial proofs. Companion to blockchain-leg (stablecoin→Treasury + Fairshake), altcoin-lens (utility vs accrual — a different axis), spec-blockchain-ecosystem (foundations/infra), spec-crypto-collapse-cluster (2022 cascade), spec-exchanges-asia.

Discipline. This page does not sort "good actors" vs "bad actors." It sorts on public-document flags: criminal convictions, civil-fraud judgments, OFAC listings, bankruptcies/insolvencies, open-source vs custodial, US-public-company status, documented government/defense facts (contracts, council seats, state stablecoins, PAC funding), and explicit none documented. Pedigree, VC backing, and council adjacency are recorded as facts. Intent is not inferred. Composition guard: a foundation is not its token holders; an exchange is not its listed assets.

The live UI is the Blockchain tab (docs/blockchain.html).

Quantum readiness (third axis)

Public-record flags and altcoin-lens utility grades do not say whether a chain still signs with ECDSA. That is a separate, load-bearing fact: Bitcoin, Ethereum, Solana, XRP, and most L2s authorize spend with ECC (harvest-now / forge-later). The tab joins every chain we can match against the Blockchain Quantum Readiness Index (qrindex.org), scraped on each fetch_qri.py run.

There is no stable API. HTML data-* attributes, visible table cells, and per-project llms.txt are parsed defensively (models/graph/qri_parse.py). A Scrapy spider (qri_spider.py) is used when scrapy is importable; otherwise stdlib + curl (macOS cert store). JSON paths on the site, if they appear, are not a contract.

QRL is the reference Stage-4 L1 in this corpus: mandatory XMSS (NIST SP 800-208) since genesis, QRI 98.5 / Stage 4 (evaluated 2026-08-20, medium confidence, draft). Mochimo, Tidecoin, and Abelian also sit at Stage 4 on the same index. That is cryptographic-readiness, not adoption. altcoin-lens still grades QRL speculative-niche on token-accrual — both can be true.

Fetcher: python3 models/graph/fetch_qri.py → data/qri_index.json (plus data/qri_raw/index.html snapshot). Prefers Scrapy when importable; otherwise stdlib HTML + curl. Cache is kept if the scrape fails. Index is AI-assisted and pre-release; scores are not proofs.

CEX/DEX rows usually have no venue-level QRI. The tab then shows host-chain scores (BTC/ETH for a CEX; Solana for Jupiter/Raydium; TRON for SunSwap). That is listed-asset spend-auth exposure, not a cleanliness rating of the exchange.

Slug map is best-effort (XRP Ledger is xrp not ripple; BNB Chain is bnb; TON is toncoin). When qrindex changes slugs, update MAP in fetch_qri.py.

Why a tab (the map gap)

The bubble map had a crypto colour bucket and a blockchain-leg (GENIUS rail + Fairshake cycle) but:

1. Bitcoin the protocol was not a node. Only Bitcoin_Strategic_Reserve, Bitcoin_Policy_Institute, and IBIT existed — policy/ETF wrappers with nothing to point at.
2. Bitfinex and Bybit were sector other (mixing).
3. Major L1s/L2s/CEXes sitting in prose (spec-blockchain-ecosystem) were not bubbles: Cardano, Polkadot, Aptos, TON, Base, Arbitrum, Optimism, OKX, dYdX, Curve.
4. Definitive adjudicated schemes marketed as coins (OneCoin, BitConnect, PlusToken) were absent, so the 2022-cascade frauds looked like the whole "scam" set.

What the flags mean (objective)

Government/defense is a typed note, not a boolean "tied to the Pentagon":

• state_policy — e.g. US Bitcoin reserve, Wyoming WYST via LayerZero, El Salvador legal tender
• council_seat — e.g. Boeing on Hedera's Governing Council (not a DoD contract)
• institutional_rail — SWIFT/DTCC/BlackRock tokenized funds
• political_pac — Fairshake funders (Coinbase, Ripple, Uniswap Labs, a16z, Jump)
• ofac / enforcement — sanctions or live DOJ cases
• none_documented — searched this corpus; no defense contract found

Closest defense adjacency in the corpus: Hedera council includes Boeing. That is labeled council_seat. No public L1 in this registry has a documented US DoD procurement for the chain itself.

Keep separate (not aliases)

Venue coverage (expanded 2026-08-22)

The first draft only had the household-name CEXes. The registry now carries 93 CEX and 54 DEX/venue rows: CMC/CoinGecko/CryptoRank 2026 volume tables (Binance, OKX, Bybit, MEXC, Gate, Bitget, KuCoin, HTX, Crypto.com, Upbit, LBank, Toobit, BingX, WhiteBIT, CoinW, Bitunix, Zoomex, Hyperliquid as a DEX), plus regional KR/JP/IN/SE Asia/LatAm/Africa/MENA/TR/AU/CA on-ramps, US institutional (EDX, Bakkt, itBit), and the dead/sanctioned set (QuadrigaCX, Cryptopia, BTC-e, Thodex, Garantex, Bittrex wind-down, FTX.US, Zipmex, Vauld, Liquid, Paxful).

Volume is not a cleanliness score. CMC's own 11-venue tape is ~85% top-five; wash-trading on some alt venues is a known measurement problem and is not turned into a flag here.

Name collisions kept un-aliased: Stargate_Finance (LayerZero bridge) ≠ AI Stargate; Binance ≠ Binance_US ≠ Binance_TR; Coinbase ≠ Coinbase_International ≠ Deribit; HTX is Huobi's successor name.

CEX and DEX histories (one line each, from corpus)

• Mt. Gox — 2014 collapse; civil rehabilitation; repayments still slipping (Oct 31 2026). McCaleb founded then sold; Karpeles was CEO at failure.
• Binance — 2023 $4.3B BSA settlement; CZ plea + 2025 pardon (fine stands); MGX $2B in USD1.
• Coinbase / Kraken — US-regulated pole; SEC suits dropped 2025; Coinbase is a Fairshake funder.
• FTX — adjudicated customer-fund theft; 25-year sentence. Not a blockchain failure.
• Uniswap — non-custodial AMM; Uniswap Labs funds Fairshake; chose Wormhole over LayerZero on security.

DEX vs CEX is a custody flag, not a moral one.

Failed businesses vs adjudicated schemes

The 2022 cascade (spec-crypto-collapse-cluster) mixed leverage insolvency (3AC, Voyager) with adjudicated fraud (FTX, Celsius, Terraform). This registry splits those flags so they can be filtered apart. OneCoin / BitConnect / PlusToken are court-found schemes marketed as coins — they were never L1s in the Bitcoin/Ethereum sense.

Known-good / known-bad (why we refuse the labels)

"Good actor" in market speech usually means still operating and listed. "Bad actor" usually means someone already lost money. That collapses:

• a protocol (Bitcoin) with an exchange failure (Mt. Gox) that used it
• a BSA fine (Binance) with customer theft (FTX)
• a privacy tool under OFAC (Tornado) with a Ponzi (BitConnect)

The tab therefore exposes the flags and lets the reader sort. altcoin-lens remains the place for utility vs token-accrual grades — a separate axis, still not a character judgment.

Graph review leftovers (not in this PR)

Thin or still-missing: MakerDAO/Aave/Compound, PancakeSwap, zkSync/Starknet/Scroll, Filecoin/Arweave as tab rows (they live in the ecosystem block), Kraken leadership dossiers. Overlay edges here are directional into protocols/sinks and do not close a financial cycle (gate: structural_edges_add_no_cycle must stay true; financial SCC stays 12).

← Research index · structured data: blockchain-registry.json · blockchain-registry.md