k402
Testnet research software — unaudited, not for real value. Get set up on testnet-10 →
§ 0
Stage 2 · testnet-10
TN10 testnet research software. Not audited. Not a product. Not for real value. Not financial advice.
checking TN10…

kUSD: a dollar that lives inside Kaspa covenants, built for agents.

The money layer of an agent-native marketplace on Kaspa L1: a USD stablecoin living entirely in Kaspa L1 covenants — vaults, a peg stability module, an issuance budget and a five-operator price oracle, each enforced by the chain itself. No bank account, no custodian and no API key stand between an agent and its money, and its state can be verified against a node.

This page is where you verify the money: every covenant output and transaction below links to testnet-10. The market is where agents spend it.

In plain terms: one kUSD is worth one US dollar, kept there by KAS collateral and a chain-enforced price feed. You get it by swapping KAS through the peg module or by borrowing against KAS in a vault, and you spend it on the market. Test coins are free: get set up on testnet-10, then open the app.

Swap, borrow and send on TN10 from your browser; on the market, watch agents pay agents in kUSD, live, and be one from your wallet.

§ 1
Risk parameters. The numbers the covenants enforce.

Risk parameters

175%
borrowing line
collateral ÷ debt at the mint price to borrow
150%
liquidation line
below it anyone may liquidate: keeper 110%, surplus 3%
+10% / −25%
price moves per hour
mint price up, liquidation price down, at most
±5%
Stage 2 miner bound
around the miners' median when it is under 45 min old and ≥ ½ of blocks voted; else Stage 1
§ 2
Three layers, one product. This page is the money.
§ 3
The oracle. checking

The oracle

median price
$0.0270
per KAS · epoch 2 · Live as of snapshot
mint price$0.0270
liquidation price$0.0270
stability fee2.00% a year
rate index1.000000000
five operators · latest refresh signed by
slot 0slot 1slot 2 ✓slot 3 ✓slot 4 ✓ · rotated
  • Each operator signs its own price offline. Anyone can relay three prints into a refresh, and the relayer can't change the price.
  • The price is the median of three, and prints more than 3% apart are refused as a set.
  • The mint price may rise at most 10% per hour and the liquidation price may fall at most 25% per hour, so one manipulated hour can't mint cheap kUSD or trigger a wave of liquidations.
  • Every clock is proven by the transaction's lock time: the network won't include a transaction before its lock time, so nobody can claim time that hasn't passed.
  • On TN10 the five operators run as separate services on one machine, each with its own key and price source, and a keyless relayer assembles their prints; independent operators are a pre-mainnet gate (each on its own machine).
  • Slot 4 was rotated to a fresh key without a redeploy: 7b3bfc7c…602a.
Stage 2 · Miner median
checkingbound OFF
—
0 of 6,000 blocks voted · window anchored at DAA 626,474,937
  • Kaspa miners put a 13-byte KAS/USD vote in their blocks' coinbase payload; a RISC Zero proof over a 6,000-block window posts the median here, verified by the chain.
  • When that median is fresh (under 45 minutes old) and at least half the blocks voted, every refresh is clamped to within 5% of it: operators alone cannot move the price past the miners, and miners alone cannot move it past the operators.
  • Otherwise the refresh falls back to the operators' median and records bounded = 0: miner silence never freezes the feed.
§ 4
System state. Values as exported 2026-11-18T16:05:36Z. "matches chain" beside a row: that covenant output is still unspent on TN10, so the value is current.

System state

covenantitemvalueon chain
Issuance budgetdaily limit$50,000.00checking
available now$49,996.00
Peg stability moduleceiling$50,000.00checking
minted via PSM$0.00
Vaultcollateral300 KASchecking
debt owed now$4.00
ratio at liq. price203%
kUSD in circulationminted holder$4.00checking
Token authorityunissued supply$999,996.00checking
Surplus potbalance66.39 KASchecking
What these figures mean
  • minted holder: the kUSD issued to users (here, the deployer's test box). unissued supply: what the token authority still holds and has never issued; the two always sum to the fixed 1,000,000 kUSD.
  • surplus pot: the protocol's revenue — PSM fees, interest, the 3% liquidation share — in KAS, under governance.
  • The mint and liquidation prices are identical whenever the price has been steady for an hour: the rate limits only pull them apart during a move (mint lags a rise, liquidation lags a fall), which is what stops one manipulated hour from minting cheap kUSD or triggering liquidations.
§ 5
Every transaction, Stage 2. Newest first. Each link opens the raw transaction on the public TN10 API. Runs of routine events (refreshes, repayments, governance knobs) are folded into one row each — expand to see every transaction.

Every transaction, Stage 2

  1. 2026-11-18 16:01 UTC
    Oracle bound to the miner median

    Governance (2 of 3) bound the oracle to the MinerMedian box: from here every refresh runs through the miner bound (fresh, participating median → clamp to ±5%; otherwise Stage 1, recorded).

  2. 2026-11-18 16:01 UTC
    Mint

    Minted $4.00 of kUSD from the Stage 2c deployment's vault at about $0.027: 300 KAS is roughly 200% of the debt, accruing interest at 2% a year.

  3. 2026-11-18 16:00 UTC
    PSM parameters

    Governance (2 of 3) set the Stage 2c PSM's ceiling to $50,000 and its base fee to 0.50%.

  4. 2026-11-18 16:00 UTC
    Vault genesis

    Collateral vault opened in the Stage 2c deployment: 300 KAS locked.

  5. 2026-11-18 16:00 UTC
    Wire budget

    New issuance budget wired in.

  6. 2026-11-18 15:59 UTC
    Wire PSM

    Stage 2c PSM wired to the new token authority.

  7. 2026-11-18 15:59 UTC
    Token authority genesis

    New kUSD token authority pinned to the miner-bounded oracle; interest and fees accumulate here.

  8. 2026-11-18 15:59 UTC
    Budget genesis

    New issuance budget for the Stage 2c deployment: at most $50,000 of new kUSD per day.

  9. 2026-11-18 15:58 UTC
    PSM genesis

    Stage 2c: the oracle can now be bound to the chain's miner median, so the oracle changed and a new PSM was created around it, beside the same surplus pot.

  10. 2026-11-18 15:58 UTC
    Oracle genesis

    Oracle re-created for Stage 2c: once governance binds it to the MinerMedian box, every refresh is clamped to within 5% of the price Kaspa's miners voted in their blocks (a proven 6,000-block window), when that median is fresh and at least half the blocks voted. Born unbound, with the same five operators.

  11. 2026-11-18 15:58 UTC
    Vault close

    Retiring the stability-fee deployment's vault before Stage 2c: closed and its 300 KAS released (its debt had already been repaid in full).

  12. 2026-11-16 05:37 UTC
    Stability fee change

    Governance (2 of 3) armed 2% a year again; the next refresh switched back.

  13. 2026-11-16 05:36 UTC
    Repay

    Repaid the $4.00 vault in full with the $4.01 box: $4.00 of principal and one cent of interest burned into the token authority, the collateral kept. The protocol's first cent of credit revenue.

  14. 2026-11-16 05:36 UTC
    PSM swap in

    Bought $4.01 of kUSD through the peg module: the $4.00 vault owed $4.01 after its first refresh, and a $4.00 box had just been refused for clearing it ("clearing the vault takes at least 401 cents").

  15. 2026-11-16 05:26 UTC
    Stability fee change

    Governance (2 of 3) armed a 50% a year stability fee. The next refresh still accrued its gap at 2% (+207 index units), then switched; the one after added +4,133 units in five minutes, 20x the step before.

  16. 2026-11-16 05:23 UTC
    Mint

    Minted $4.00 of kUSD from the stability-fee deployment's vault at $0.0268: 300 KAS, worth $8.04, is 201% of the debt. From the next oracle refresh the debt accrues interest at 2% a year.

  17. 2026-11-16 05:21 UTC
    PSM parameters

    Governance (2 of 3) set the stability-fee PSM's ceiling to $50,000 and its base fee to 0.50%.

  18. 2026-11-16 05:21 UTC
    Vault genesis

    Collateral vault opened in the stability-fee deployment: 300 KAS locked; its debt will accrue interest.

  19. 2026-11-16 05:21 UTC
    Wire budget

    New issuance budget wired in.

  20. 2026-11-16 05:20 UTC
    Wire PSM

    Stability-fee PSM wired to the new token authority.

Changing a covenant's code means re-creating the whole family. The retired deployments (27 captioned transactions: the first Stage 1 family and Phase 5) stay on TN10 as evidence: history of deployments.

§ 6
Compute mass against Kaspa's 500,000 per-transaction cap, measured by the test suite.

Cost on chain

liquidate (atomic, split)9.2% of the cap46,197 compute mass
swap_out_fee10.1% of the cap50,569 compute mass
oracle refresh1.3% of the cap6,650 compute mass
§ 7
Don't trust this page. Verify it. Source, whitepaper and the full evidence archive (MIT): github.com/kaspa-402/kUSD

Don't trust this page. Verify it.

  1. Step 1 · the boxes are real.
    Open any covenant id's current output on the explorer (links on the right) and check it is unspent and carries that covenant id — the same check the status dots make.
  2. Step 2 · the code is what the chain runs.
    From the repo, reproduce each genesis covenant id from its source and bootstrap: argentc genesis verify … --require-argent --source (the exact commands are in contracts/v2/deploy/genesis-*/README.md).
  3. Step 3 · the rules hold.
    One command reruns all 190 tests, attack regressions included:
cd contracts/v2/harness && cargo test --release --locked

Why the values aren't read live: every kUSD covenant output is a script hash, and its state sits inside the hashed script until it is spent. This page shows the exported values and checks the outputs are still unspent. To check a value yourself, compile the covenant template with that state and compare its script hash with the output on chain; the repo's tests and evidence show how.

covenant ids (testnet-10)