B1 (Build B milestone 1): TRC20 hashed-timelock contract on Tron, the
Tron leg of the BCH<->Tron atomic swap. Validated on Nile testnet, all 6
checks PASS:
- claim with correct preimage succeeds, credits receiver
- wrong preimage reverts at the sha256 guard ("bad preimage")
- refund reverts before timelock ("not expired"), succeeds after
Uses SHA-256 (matches AtomicSwapHTLC.cash) so one preimage spans both legs
in B2. Compiled solc 0.8.35 evmVersion=istanbul for TVM; swap id computed
client-side via keccak256(abi.encodePacked(...)) to mirror the contract.
contracts/tron/TronHTLC.sol, TestTRC20.sol
scripts/tron-htlc-b1.mjs (harness), tron-balance.mjs (Nile balance util)
Also folds DigiByte (DGB) into the chain landscape: List A native-Script
HTLC (reuses BTC P2WSH HTLC code, ~15s blocks) + dgbDGB row in List B2.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Build B scope: both flavors (trustless swap + lock-and-mint tUSDt), a
6-milestone test program (B1 Tron HTLC -> B6 bonded trust floor), deps,
locked decisions (sha256 both legs, oracle witness v1, USDm swap counterpart),
and honest trust-minimized caveat for the bridge. Adds reusable
tron-gen-address.mjs (Nile keypair, gitignored).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Two-list landscape for the multichain taxonomy cells:
- List A (DEX-Atomic): trustless swap, value leaves source; grouped by
primitive (native Script HTLC / contract HTLC / XRPL crypto-condition /
Monero adaptor-sig).
- List B (ASM-Multichain): lock-and-mint IOU onto BCH, value stays;
backend pipeline + two peg types (stablecoin-backed vs native-coin 1:1)
+ B1 stablecoin sources and B2 native L1 coins (BTC/ETH/SOL/BNB/TRX/...).
Adds 2026 stablecoin distribution data, build priorities (USDT-Tron first),
and positioning: BCH as a decentralized, no-KYC interchain settlement hub
(honest trust-minimized caveat for the List B bridge). README cross-linked.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Captures the conceptual material from the latest discussion (not in the existing T6 docs):
- enforcement: BTC never compares states; staleness = revocation-key handover; the victim
delivers the revocation sig in their own penalty tx; the watchtower bot is a relayer, not
an oracle (no trust); CSV window ~1 day; an honest close can't become stale.
- receive model: LN sign-to-receive = "be present" (like cash) -> can't receive offline;
T6 inherits it; native BCH receives while offline (the digital ideal restored).
- topology: chain-as-backbone + light clients vs LN's node-mesh-with-hubs; the chain carries
the always-on burden, not nodes; custodial services can run T6 without a node.
- multi-party: N-of-N funding easy, N-party live operation hard (all must sign); channel
factories; the BCH board extends to N trivially.
- tokens: T6 uses an NFT only (no fungible token); value stays BTC, BCH holds only the record.
- economics: custodial-LN fees vs BCH (sourced) — spread is a fiat markup not an LN haircut;
entry/exit is on-chain (WoS 1.95% / Cash App up to $20 vs BCH <$0.001); real cost is custody.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Standalone "why" companion to the technical T6 docs. Captures the reasoning thread:
- the epistemics (finite-time expertise -> specialized truth doesn't travel by consensus;
lacks-expertise != wrong, but deferral propagates error in a captured environment; the
honest form is the derivation, not a borrowed "90%");
- the information environment (Vol.13/14: the cash->"digital gold" reframe + the protocol
track; Bezmenov four-stage lens; the demoralization finding that facts don't move a
conditioned mind, grounded in motivated reasoning; Alinsky tactics; media capture);
- why BTC users feel insecure (HODL/digital-gold conditioning makes holding feel safe and
use/swap feel like loss/heresy -> they accept lock-and-wait models like LN);
- the conclusion: reasoning fails on a conditioned mind, so demonstrate -- show don't accuse,
meet them in their experience (Model 3 / T6 keep-your-BTC on-ramps), demos are
uncounterattackable because they need no intentionality claim;
- strategy and integrity align (lenses, not accusations).
Discipline: documented mechanism vs interpretation kept separate throughout.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Self-contained explainer of the T6 model: what it is (a BTC channel whose state record
lives on the BCH chain instead of an LN node), how it works (clerk/bailiff — BCH records,
BTC settles), the step-by-step mechanics (open / transact / cooperative close / unilateral
exit + revocation penalty), how it differs from LN (state-record location → no node, no
lost-state, permissionless watchtower, two light wallets), the trust model (penalty as a
"BTC tax"; denial > penalty on BCH-native), LN interop (HTLC-capable T6 = LN channel with a
BCH watchtower; submarine swap to fund), what's validated on real coins, and why it matters.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- scripts/chipnet-channel-board-multi.mjs — T6 scale test: 3 concurrent channels = 8
addresses (6 sign-only/unfunded parties + 1 permissionless watchtower + 1 operator),
validated on chipnet. The board side scales freely; parties only sign.
- complete record §7: signing is NOT an on-chain event (LN carries signed-but-unbroadcast
BTC txs); T6 = two light self-custodial wallets vs LN = light wallet + a persistent node
(BCH absorbs the node's state-keeping + watchtower jobs); LN↔T6 interop — an HTLC-capable
T6 channel is an LN channel with a BCH watchtower, so LN payments route in off-chain;
funding from LN funds crosses via on-chain BTC (submarine swap preferred).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
T6c: the no-cooperation Lightning-faithful layer. The commitment pays the closer to a
revocation output [ OP_IF <revPub> CHECKSIG OP_ELSE <CSV> CSV DROP <closerPub> CHECKSIG
OP_ENDIF ]. Honest close = ELSE after the CSV delay; a STALE broadcast = the counterparty
sweeps via the revocation key (IF branch), immediately — the penalty.
Validated on real signet (scripts/btc-signet-penalty.mjs): funded 2-of-2 (6163...), Alice
broadcast a revoked commitment (b76d17..., her 4500 to the revocation output), Bob penalised
it (b35186..., seized the 4500). Cheating costs the whole balance -> unilateral exit is
trustless. BTC enforces the penalty (it holds the funds); the BCH board says WHEN to act.
So T6 is complete: BCH records state, BTC locks/settles/penalises. Only the honest CSV-
delayed self-close (liveness mirror) is unrun. Record + tracker updated.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
T6 ("Bitcoin Cash as a better Lightning", for LN users): keep LN's model, replace LN's
private must-not-lose must-stay-online node with a public on-chain BCH record.
- contracts/BchChannelBoard.cash + scripts/chipnet-channel-board-test.mjs — the BCH board:
both-signed, strictly-monotonic on-chain state record; stale update rejected; anyone reads
it; impossible to lose. Validated on chipnet (state 0->1->2).
- scripts/btc-signet-channel.mjs — the BTC leg on REAL signet (bitcoinjs-lib v7): funded a
2-of-2 (764238..., 10000 sat), cooperative close (af4db7..., Alice 5700 / Bob 3800).
Chained off the unconfirmed funding — open->settle in one pass. BCH records, BTC settles.
Complete design record: Parameters/T6_BitcoinCash_LN/T6_BTC-BCH_complete-record.md
(the full BTC<->BCH spine, the 3-model menu, every T-model with results). Tracker updated
(T5/round-trip/Model-3/T6 all validated).
Pending: T6c trustless unilateral-exit (penalty/revocation on BTC) — the no-cooperation
Lightning-faithful layer. package.json left uncommitted (BTC deps mixed with another session).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Hardening pass on the USDt→τUSDt wrap (builds on 0170's end-to-end validation).
Security fix — found via robustness testing:
- UsdtLockNile used require(token.transfer(...)) assuming a bool return; real
Tether-style USDt returns nothing → the refund path reverted, which would
STRAND the locker's funds. Fixed with SafeERC20 handling (no-data or true) on
both transfer/transferFrom; redeployed to Nile (TVahBUBqUDLqVjDn77Yav…Vnn2o).
- scripts/tron-wrap-robustness.cjs: R2 (too-early) and R3 (consumed) revert for
the right reason (decoded on-chain revert). R1 (self-refund) validated against
a clean standard token — contracts/tron/MockUsdt.sol + tron-refund-validate.cjs
— since the Nile faucet USDT has a non-standard transfer that reverts for
contract senders.
Naming model (revises the prefix-primary stance):
- identity = CashToken category; human name = USDt + a network badge (the "BIC");
the ε/τ prefix is a text-only fallback. Applied to the mockup (ASM·Multichain:
name+badge, prefix in button/stepper, synced selectors, new source-chain
dropdown Ethereum/Tron→ε/τ) and to ASM-Multichain.md / README.md, plus the
cross-chain-redemption non-goal (per-chain IOUs + AMM swap, not a universal IOU).
Design record:
- Parameters/ASM DEX/cross-chain-asm-concept-and-state.md — concept, the
trustless-vs-authority analysis, oracle vs light-client vs zk, the VM wall, the
reference hacks, the trust ladder, and the 48-hour build state.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Build A — DEX, validated on chipnet:
- scripts/chipnet-amm-usdm-pusd.mjs: USDm/PUSD constant-product swap on the
first-party SharePairAMM (ADR-0030); price discovery 1.000→1.440 PUSD/USDm,
constant product held. No Cauldron needed (it's mainnet-only).
Build B — cross-chain wrap, validated end-to-end across Tron Nile + chipnet:
- contracts/DeviantWrapMintIOU.cash: mint-on-attestation IOU covenant —
checkDataSig oracle gate, NO BCH backing (dust only; value is the source
lock, ADR-0035), replay-bound to the reservoir outpoint.
- scripts/chipnet-wrapmint-iou.mjs: BCH leg — genesis reservoir into the
covenant, oracle-attested mint of τUSDt, replay rejected by the covenant.
- contracts/tron/UsdtLockNile.sol + scripts/tron-deploy-lock.cjs: Nile lock
contract (lock/markConsumed/refund), deployed at TXmMRnau9Ct…ZintH.
- scripts/tron-bch-wrap.mjs: the oracle watcher — locked 0.5 real USDt on
Nile → minted 500000 τUSDt on chipnet → markConsumed. Genuine USDt→τUSDt.
Docs (Parameters/ASM DEX/ASM-Multichain.md): which source chains qualify,
reverse unwrap, three-tier trust model, source-chain prefix marker, and the
cross-chain-redemption design decision (universal IOU = non-goal; per-chain
IOUs + BCH AMM swap = the answer; CCTP does not cover BCH).
Deps: tronweb ^6.3.0, solc ^0.8.35. Keys/state gitignored.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Record the decision: the fiat on/off-ramp is built as an independent product
(own repo/deployment/brand); the Deviant wallet is one consumer via API/SDK,
alongside third-party embedders. Follows the established split-out + white-label
pattern (CoinSpectrum, Deviant Labs co-brand). ADR-0036 on-ramp aggregator
becomes a client of the product.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Adds Parameters/DEX_P2P/fiat-ramp-product.md — the product layer over the
settlement engine in p2p-fiat-crypto-exchange.md:
- Non-custodial infrastructure: catalogue + escrow + oracle + reputation;
platform holds nothing and converts nothing (light-touch posture).
- Two counterparty classes on one backend: individuals (pure P2P, non-KYC,
self-proved zkTLS) and specialized operators (licensed/vetted, automated,
carry own compliance) — with class + jurisdiction disclosure.
- Trade flows for each lane; regulatory posture (operators carry compliance;
regulation follows operator AND customer; geo-controls + disclosure).
- Positioning: "BestChange's network + reach, but escrow- and oracle-protected";
micro-swap wedge; Telegram + web distribution.
- Scope: fiat↔crypto only; crypto↔crypto is the separate multichain DEX.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Capture the DEX/P2P design worked out this session under Parameters/DEX_P2P/:
- dex-architecture.md — the exchange model (AMM is an order type, not a venue);
one infrastructure with two entry ramps + two skins; THORChain-insolvency
note validating the no-protocol-vaults thesis.
- bch-dex-development.md — native BCH order-book engineering spec (order
covenants, multi-input taker settlement, the 100KB tx cap, indexer/matcher).
- cross-chain-htlc-ramp.md — HTLC entry-ramp spec (two-leg atomic swap,
T_B<T_A timelock ordering, BCH covenant, BTC/EVM counterparty legs).
- p2p-fiat-crypto-exchange.md — fiat-leg payment oracle + settlement covenant;
receipt-vs-sender-proof decision resolved to sender-proof (2b) via the
three-adversary test (new §4a), refund-default victim mislabel fixed.
Design records only; no code.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Follow-up cleanup to the USDm -> FIATm reorg: update the remaining cross-references
that still pointed at Parameters/usdm_complete-design_FOR-design-record.md to its
new path Parameters/FIATm/fiatm_complete-design_FOR-design-record.md.
Touched: src/oracleUsdm.ts, wiki/decisions.md, wiki/sound-money-primer.md, and the
FIATm/MBA-Shares design records. Adds Parameters/FIATm/fiatm-comparisons.html.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Brings the whole project into git (previously only the on-ramp feature and
a few parallel files were tracked). Includes: wallet core (src/), CashScript
contracts/, the Capacitor shell (shells/), scripts/, tests, and the docs
corpus (wiki/, snapshots/, Parameters/, Archive/, KnowledgeBase text).
Excluded via .gitignore: node_modules/, dist/, .env (secrets), testnet WIF
key files (.chipnet-*.json, devWallets.ts), KnowledgeBase/RAW media (~9.6 GB),
transcripts/ (session logs), and machine-local config (.claude/*.local,
*.lock). Secret-scanned before commit — no credentials in the tree.
Note: src/swapMath.ts, src/swapBuilder.ts, contracts/OpenMintToken.cash are
legacy (OP_RETURN era) but still imported; committed as-is, flagged for a
separate removal once de-referenced.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
