Battle test: PegLab vs Parker’s GitHub PoC

Parker’s public model is not a dollar printer. It is kaspa-explained (Testnet-10 receipts, wrap lab, Sprout Harbor) plus tn12-covenant-vault-demo (accepted TN12 vault/escrow/assurance proofs).

Authority: DOCTRINE.md. This file scores this PegLab checkout against that model. PegLab is allowed to lose. If a row says Parker wins, do not “fix” it by calling tPEG USD.

TESTNET TOY. NOT USD. DO NOT RAISE AGAINST tPEG.

Sources dated September 2026: Parker README + docs/wrap-poc.md + docs/backed-receipt.md; PegLab src/engine.mjs, MAINNET.md, VISION.md.

1. What Parker actually proved

Native receipt (the object PegLab should become)

From backed-receipt.md / wrap lab:

One receipt unit equals one locked sompi. Not USD.
Transfer / split / merge conserve quantity and backing.
Redeem pays the holder’s P2PK. No mint refill.
Sponsor input pays fees. Principal cannot be skimmed.
Identity is genesis outpoint + covenant ID + template + series. A name is not authenticity.
Live TN10 evidence exists for create, split, merge, move, partial and full redeem (accepted txids in Parker’s public review docs).

Wrap / “stable” path (explicitly not money)

Lock 100 pUSD (freely minted Sepolia test token) → issue Kaspa units → burn → release.
Trusted local oracle. Kaspa does not verify Ethereum consensus.
wTestUSD cannot buy the town’s crops. Written in the README.
Round trip recorded 7 September 2026; vault and claim balances ended at zero.

Economy (Sprout Harbor)

Real tKAS spends with rules: wage UTXO, vouchers, pledges, schedule check.
21 accepted TN10 txs for one order; reload verified 7 September 2026.
Game art is interpretation; signatures and spend limits are on-chain.

TN12 covenant lab

Accepted: delayed vault, pledge release/refund, buyer escrow, batch-assurance, payload receipts, adversarial rejects.
Vocabulary: SCRIPT_ENFORCED vs PLANNER_ONLY vs INDEXER_DERIVED vs WALLET_POLICY.
Honest gaps: no AMM, no lending, no production oracle, no user-wallet signing on the accepted proofs (local keys), no independent audit, mainnet out of scope.

Parker’s model in one sentence: lock real tKAS (or a named test wrap), prove the spend rule, journal the txid, never call it a bank dollar.

2. Scorecard

Test	Parker PoC	PegLab (this repo)	Winner
What the unit is	1 receipt = 1 sompi locked in the output	tPEG minted at an admin oracle (default 100_000 sompi/tPEG) plus a 2 tKAS constant-product pool	Parker for money. PegLab is a different object: a depeg toy.
Dollar claim	Forbidden. wTestUSD cannot buy crops. pUSD is freely minted test.	Forbidden in copy (WILL DEPEG). Engine still has a USD-shaped oracle + pool	Tie on copy. Parker wins on mechanism (no dollar feed).
Redemption	Holder spends backing; no oracle	Redeem tPEG for tKAS at admin price	Parker
Fees	Sponsor input; no skim of principal	Lab UI does not yet enforce sponsor-fee on a live UNIT tx	Parker
On-chain evidence	Accepted TN10/TN12 txids, JSON journals, explorer links	Engine + tests + optional genesis submit; no public accepted UNIT transfer/redeem journal	Parker
Caps	10 tKAS teaching bound on receipts	10 tKAS backing, 10k tPEG, 2 tKAS pool — same teaching scale	Tie
Identity	Genesis / template / series; name ≠ authenticity	Same sentence in PegLab copy	Tie
Fail closed	Wrong signer/amount/skim rejected; wrap rejects duplicate deposit/burn	Engine rejects ALIGN, OVER_CAP, PAUSED, STALE_ORACLE, BAD_ADMIN, TINY_POOL	Tie on intent. Parker has VM + node rejects on record.
Oracle honesty	Wrap oracle is named as trust. Native receipt has none.	Admin key is the price. Lab shows the depeg on purpose	PegLab as a lesson. Parker as a product.
AMM / “peg defense”	Explicitly no AMM on TN12 (“current gaps”)	Tiny pool is the depeg demo	PegLab for teaching. Parker for not pretending a pool is a peg.
Timeout / escrow	TN12 buyer escrow, delayed vault, pledge refund — accepted	Documented for KaChat (`KaChatPayTimeout`); not live	Parker
Indexer / replay	Payload receipts, replay guards, claim classes	No UNIT indexer row yet	Parker
Wallets	Encrypted browser journal; wrap keys stay `.local/`; Kaspa does not get a public signing server	KasWare header (Gramlane `getBalance`) + localhost faucet; public site has no host key	Tie (different surfaces). Parker is ahead on recorded independent-role txs.
Product scene	Town that spends tKAS with rules	KaChat-on-phone vision; PC told KaChat is mobile-only	PegLab on consumer story. Parker on executed economy.
Funding / law map	Wrap-poc “from here to real capital” table (asset, oracle, users, supply, finality, recovery, assurance, obligations)	MAINNET.md + STABLES-GUIDE.md: Maker/Circle/Liquity funding, GENIUS, never raise tPEG	PegLab on capital honesty. Parker’s table is tighter on ops.
Mainnet	Out of scope; readiness doc is gaps	Out of scope as money; successor A = Parker’s receipt	*Parker already is* successor A**

Verdict: PegLab does not beat Parker as a Kaspa unit. Parker’s backed receipt is the thing MAINNET.md calls successor A. PegLab wins only as the public depeg classroom and the fundraising/KaChat map. If you deploy PegLab’s pool+oracle on kaspa:, you lose the battle Parker already won in copy: *do not let a test wrap buy the town.*

3. Battle on Parker’s own tests

Run PegLab’s engine against Parker’s receipt rules. Failures are expected where PegLab is a different machine.

Parker rule	PegLab today	Result
Quantity = backing sompi	Mint: tKAS in must be a multiple of oracle price, not 1:1 sompi	FAIL as receipt. PASS as “toy mint.”
No price oracle on the unit	`DEFAULT_ORACLE = 100_000n` sompi/tPEG; mint/swap require `oracleLive`	FAIL as receipt. PASS as depeg lab.
Sponsor pays fees; no skim	Not encoded in the live UNIT path	FAIL until successor copies Parker’s sponsor pattern
Transfer conserves units	`transfer` conserves tPEG quantities	PASS (tPEG, not sompi)
Redeem to holder P2PK	Redeem burns tPEG, pays tKAS at oracle	FAIL as 1:1 sompi redeem. Different payout.
Tiny pool cannot defend $1	`cannotDefendPeg` true at seed; depeg lab moves pool >5%	PASS — this is PegLab’s job
Name ≠ authenticity	Copy + series id	PASS in docs
Accepted txid journal	Missing for UNIT	FAIL
wTestUSD cannot buy crops	tPEG must not sit in KaChat as money	PASS in KACHAT.md; FAIL if anyone lists lab tPEG in 4.x

4. What PegLab should steal from Parker (copy-paste)

Not the town art. The enforcement.

Successor A = Parker’s receipt. Kill the admin USD oracle on any asset KaChat would show. 1 unit = 1 sompi.
Sponsor fee input. Parker already wrote it. MAINNET.md already requires it for 0-KAS receivers.
Evidence pack. Every “done” is an accepted txid JSON like Parker’s wrap verification record — not a screenshot of the depeg UI.
Claim vocabulary. Label PegLab engine rows ENGINE_SPEC, SilverScript SCRIPT_ENFORCED, homepage copy COPY_ONLY. Do not mix.
Wrap honesty. If you ever host USDC, say “trusted issuer/bridge” the way Parker says “trusted test oracle.” wTestUSD cannot buy crops becomes tPEG cannot pay KaChat.
TN12 primitives for KaChat. Timeout pay = Parker’s delayed vault + escrow + pledge refund. Do not invent a third covenant family.
.local/ never on the public host. Parker’s wrap signing server is loopback-only. PegLab faucet already is. Keep it.

5. What Parker should steal from PegLab (if anything)

Depeg as a first-class demo. Parker refuses a fake peg. PegLab *shows* why a 2 tKAS pool dies. That lesson belongs next to wrap lab, not instead of receipts.
KaChat / phone / “Log in”. Parker’s proofs still used local keys; TN12 defers user-wallet signing. PegLab wired Gramlane KasWare getBalance on the public page.
Capital instruments. Parker’s “from here to real capital” is ops. PegLab’s Maker (MKR private sales, no DAI ICO), Circle (equity → IPO, GENIUS), Liquity (company + immutable core) table is how you *pay* for successor C/D without a sister token.

Do not merge the labs into one ticker.

6. Outcome

If you wanted…	Winner	Next move
A Kaspa-native unit people can transfer and redeem	Parker receipt	Stop minting tPEG at an oracle. Implement successor A on TN10 with Parker’s sponsor-fee rules. Journal txids.
A dollar	Neither	Circle/Tether or a licensed company. Parker already refused. PegLab MAINNET.md same.
A teaching depeg	PegLab	Keep WILL DEPEG. Do not mainnet the pool.
Chat settlement	Neither live	Parker has escrow/timeout on TN12. PegLab has the KaChat story. Combine: Parker primitive + KaChat signer.
Something to raise money against	Neither	Grants for indexer/signer. Equity only if you are an issuer.

Short version: Parker’s GitHub PoC already *is* the honest Kaspa unit. PegLab is the honest warning that a pool is not a peg. Battle-test score: receipts 1–0 Parker; classroom 1–0 PegLab; dollars 0–0. Deploy the receipt. Keep the lab. Do not put PegLab’s oracle on mainnet.

Parker: https://github.com/parker2017code/kaspa-explained · https://github.com/parker2017code/tn12-covenant-vault-demo PegLab: this repo · MAINNET.md