Examples — trading & market data (11–18)
11. Two-phase transaction confirmation (MCP)
// Step 1 — preview (executes nothing, binds the args to a hash)
{ "tool": "furlpay_preview_execution",
"args": { "tool": "furlpay_trade",
"args": { "side": "buy", "asset": "USDC", "amount": 25, "price": 1.00 } } }
// → { executionHash: "…", netEffect: … }
// Step 2 — confirm with the hash, within 10 minutes, exactly once
{ "tool": "furlpay_confirm_execution", "args": { "executionHash": "<from step 1>" } }Semantics: the policy gate re-runs on the stored amount at confirm time; the hash is single-use; attempts and refusals both land in ~/.furlpay/mcp-audit.jsonl (args hash only, never raw arguments).
12. x402 payment challenge (SIMULATED)
furlpay x402 pay \
--challenge-header 'WWW-Authenticate: x402 recipient="GBLX99API", amount="0.0005", token="USDC"' \
--max-cost 0.005Status: SIMULATED / LOCAL_SIMULATION — no live facilitator exists in this build, and the response says so on every field. Replay protection runs against a file-locked nonce ledger. Nothing here moves value.
13. Stellar SDEX trading (labeled simulation)
furlpay trade limit \
--side buy \
--asset JQQQ:GDXFE4P76X667RKYHVMZ4H5XQWPLK3X72Z7YXZGHK8R5M7OQ3V7QJQQQ \
--against USDC:GA5ZSEJYB37JRC5AVCIA5MOP4RHTM335X2KGX3IHOJAPP5RE34K4KZVN \
--amount 25 \
--price 104.20Status: SIMULATED — a deterministic local simulation of Horizon submission; no transaction is broadcast. Results carry the simulated label (audit H-03: the CLI prints the simulation status rather than a checkmark). Regulated-asset trades report sep8Status: "not_performed" — compliance is never claimed.
14. Hyperliquid market data (LIVE)
furlpay hyperliquid status
furlpay hyperliquid markets
furlpay market BTC --slippage 0.5 --side buyReal /info responses from Hyperliquid validators: validator connectivity, mark price, funding, open interest, L2 depth bands, OFI and micro-price. No credentials needed — this plane is read-only.
15. Hyperliquid signed perp order (LIVE — real funds)
Risk: this places a real, signed order on Hyperliquid L1. Size it as if it were your money, because it is.
export HYPERLIQUID_PRIVATE_KEY=0x... # API wallet recommended
furlpay hyperliquid trade --symbol BTC --side buy --size 0.05 --leverage 5 --limit 75000The pre-trade risk engine runs fail-closed: asset metadata, L2 book and clearinghouse-state fetch failures refuse the order instead of skipping the guard; the USDC peg guardrail ingests USDC/USD (a feed outage is a warning, never depeg evidence); notional/turnover/margin math is scaled-BigInt. Without FURLPAY_LIVE=1 the master gate keeps the surface inert.
16. Hyperliquid shorting (LIVE — real funds)
export HYPERLIQUID_PRIVATE_KEY=0x...
furlpay hyperliquid trade --symbol BTC --side sell --size 0.05 --leverage 5
furlpay account positions # watch the short
furlpay hyperliquid close BTC # market buy-backReadiness research before entry (read-only, over your supplied data — never an execution trigger):
furlpay short signal --book depth.json --tape trades.json --bucket 1000Binance futures shorts are NOT IMPLEMENTED (declared false in the capability registry — there is no method on the surface).
17. HLP vault information (LIVE)
furlpay earn hlp statsEquity/APY/PnL derived from the live /info vault response. Fields the public API does not report (share price, open position count) print as unavailable — never fabricated benchmarks.
18. HLP deposit / withdraw (LIVE — real funds)
Risk: these are real signed
vaultTransferactions. Deposits and withdrawals carry Hyperliquid's 4-day withdrawal lockup; the CLI checks lockup eligibility client-side before signing (durable deposit timestamps — audit H-04 fix).
export HYPERLIQUID_PRIVATE_KEY=0x...
furlpay earn hlp deposit 1000
furlpay earn hlp withdraw 100Fail-closed without a configured signer: the CLI refuses rather than simulating.