The Prompting Company

Benchmark report

Task verified

6m 39s 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
Two benchmark-request schema rejections for the credential form, plus having to grep the CLI binary with `strings` to confirm the documented auth env var names.

Where the time went

6:40 agent time
  • Research0:29
  • Attempt4:17
  • Verify1:54
6m 39sTotal elapsed
2:48Time to first success
0:00Waited on the owner

Judge reason

Saved CLI commands/outputs (commands/01-15, outputs/*.stdout) show the whole round trip ran through /usr/local/bin/kraken v0.5.1 with credentials injected as KRAKEN_API_KEY/KRAKEN_API_SECRET by scripts/krun.sh; no HTTP/SDK/MCP call appears in the transcript. Pre-trade: 'kraken order cancel --cl-ord-id ctrl0' returned EOrder:Unknown order and open-orders was {}. Start balances USD 76.422 / BTC 0.0. Buy: 'order buy XBTUSD 20 --type market --oflags viqc --cl-ord-id ctrl0mrt1794466621' (viqc is a documented --oflags value in the CLI help) -> OQTPZM-OFSSH-KLBYOD, status closed, volume_executed 0.0002726, cost 20.0, fee 0.16, oflags viqc,fciq; trade THVKY5-JG2DV-TKRBPT price 73367.8. Fee charged in USD (fciq), so BTC net of fees = 0.0002726, confirmed by balance 56.262/0.0002726 (delta -20.16 USD). Sell: 0.00027260 BTC (already on the 1e-8 lot_decimals=8 increment) -> OH3IXD-RDPGJ-YDGAEV, closed, volume_executed 0.0002726, cost 20.00003, fee 0.16 USD, trade price 73367.68158. End balances USD 76.102 / BTC 0.0: net -0.320 USD (<1) and 0 BTC change (<0.50 USD). closed-orders --start 1794466600 returned count=2 and open-orders {} afterwards, so no other order was placed. Every printed price, size and fee in REPORT.md matches the saved query-orders/query-trades JSON; stderr shows 'live: this goes to the real Kraken account' and mode:live on each placement. Credential grep over the saved tree was clean and no .cmd/.stdout file contains secret material.

Attempt record

  1. 1:42Answered from saved key: Kraken Spot API credentialsno wait
  2. —First success verified

Attempt artifacts saved.