The Prompting Company

Benchmark report

Task verified

3m 49s elapsed

The saved attempt met the documented success criterion.

Experience score

96/ 100

Agent-Ready
  • SInstall
  • SAuth
  • SExecute
  • ADocs
  • SCurrent
Experience measures friction in the tested task. The verdict says whether it succeeded.
Agent profile for this run: Install S, Auth S, Execute S, Docs A, Current SSInstallSAuthSExecuteADocsSCurrent
Agent profile · this run
Judge's note · what slowed the agent most
A guessed doc URL (/reference/api/retrieve-a-session) 404'd and an auth.md fetch was proxy-refused, costing one extra fetch round before /get-a-session resolved.

Where the time went

3:49 agent time
  • Research0:18
  • Attempt2:18
  • Verify1:13
3m 49sTotal elapsed
1:40Time to first success
0:00Waited on the owner

Judge reason

Saved code (browserbase_session.py) uses only urllib against https://api.browserbase.com/v1 with the key read from a runtime file and sent solely in the X-BB-API-Key header. Create: POST /v1/sessions body {"timeout":120} -> HTTP 201, id de1558ff-11b0-4467-9b14-daba5b720a67, status RUNNING, expiresAt exactly startedAt+120s confirming the <=5min option applied (outputs/1_create_session.json). Retrieve: GET /v1/sessions/{id} -> 200 RUNNING. List: GET /v1/sessions -> 200, 229 entries; independent parse of outputs/3_list_sessions.json confirms the same id present with status RUNNING. Terminate in the finally block: POST /v1/sessions/{id} {"status":"REQUEST_RELEASE"} -> 200, status COMPLETED, endedAt 2026-11-03T04:42:30.516Z. Final GET -> 200 status COMPLETED. run_stdout.txt printed values match every saved response exactly. No Playwright/Puppeteer/Selenium import or connectUrl use; connectUrl/signingKey redacted in saved files and a grep of the home dir for credential-shaped strings found only session UUIDs. Transcript shows one live run, no mocks or hard-coded responses, CLI only.

Attempt record

  1. 1:08Answered from saved key: Browserbase API keyno wait
  2. —First success verified

Attempt artifacts saved.