The Prompting Company

Benchmark report

Task verified

8m 6s elapsed

The saved attempt met the documented success criterion.

Experience score

89/ 100

Agent-Ready
  • SInstall
  • AAuth
  • SExecute
  • ADocs
  • SCurrent
Experience measures friction in the tested task. The verdict says whether it succeeded.
Agent profile for this run: Install S, Auth A, Execute S, Docs A, Current SSInstallAAuthSExecuteADocsSCurrent
Agent profile · this run
Judge's note · what slowed the agent most
Having to grep a 1.7MB minified CLI bundle for the undocumented OKX_SECRET_KEY/OKX_PASSPHRASE/OKX_SITE env var names, since docs cover only config.toml.

Where the time went

8:06 agent time
  • Research0:25
  • Attempt5:04
  • Verify2:37
8m 6sTotal elapsed
3:20Time to first success
0:00Waited on the owner

Judge reason

Saved evidence proves a complete live OKX US spot round trip via the official okx CLI 1.5.0 only (/usr/bin/okx -> @okx_ai/okx-trade-cli), with no curl/SDK/HTTP in the 43 transcript commands. BTC-USD was shown absent on the US site (error 51001, outputs/05b), so BTC-USDC (quoteCcy USDC, lotSz 0.00000001) was used. Pre-trade: open orders [] and the ctrl0 cancel attempt printed (51400). Buy: `okx spot place --sz 20 --tgtCcy quote_ccy --clOrdId ctrl0mrt1794467501` -> ordId 4092469718241652737, state filled, avgPx 73457.9, fillSz 0.00027226 BTC, fee 0.00000095291 BTC (feeCcy BTC), net 0.00027130709; sell of exactly 0.00027130 (net floored to lotSz) -> ordId 4092470280580378625, filled, avgPx 73447.5, fee 0.069742073625 USDC. Printed figures match outputs/31,32,41,42 exactly, and balance deltas reconcile to the cent: USDC 74.8522906374585 -> 74.7092074598335 (-0.1431, <1 USD) and BTC 7.545e-9 -> 1.4635e-8 (+7.09e-9 = $0.0005, <0.50 USD). Only these two orders fall inside the session window (07:11:41Z/07:11:58Z vs session 07:08:59-07:13:44Z); history entries are from earlier sessions. Open orders after: []. Credential leak scan over the home dir against the real secret values returned hits=0; no credential appears in any saved file.

Attempt record

  1. 2:17Answered from saved key: OKX live API credentials (US site)no wait
  2. —First success verified

Attempt artifacts saved.