Visits per provider and the no-show rate for Q1
The request
The task request supplied to the adapter, alongside the workspace below. Runtime instructions, authorized tools and skills also affect the system. Containerized runs withhold host-side references; local mode is not an isolation boundary.
Can you pull our Q1 numbers out of the scheduler export - how many visits each provider did in each month, and how bad our no-show problem is? Save it as visits.xlsx with live formulas. The front desk notes say how we count them.
Workspace
- cancellation_fees_billing.csv
- front_desk_notes.txt
- provider_roster.xlsx
- visits_export_q1_2026.csv
Expected deliverables
- visits.xlsx
This task also has a task-specific grader (check.py).
Recorded acceptance checks
The table uses check names and outcomes from the frozen campaign ledger. Acceptance requires all required predicates to pass; a predicate may itself contain a tolerance or coverage threshold. Its scope is part of the measurement contract.
| Recorded check | Type | Required | Failed / observed |
|---|---|---|---|
| visits.xlsx exists | file_exists | yes | 0 / 6 |
| live formulas | xlsx_has_formulas | yes | 0 / 6 |
| no error cells | xlsx_no_errors | yes | 1 / 6 |
| merged provider February visits | xlsx_value_present | yes | 1 / 6 |
| merged provider quarter visits | xlsx_value_present | yes | 1 / 6 |
| second provider no-show count | xlsx_value_present | yes | 1 / 6 |
| clinic total visits | xlsx_value_present | yes | 3 / 6 |
| no-show rates per provider | custom | yes | 1 / 6 |
Current task declaration (8 checks)
From task.yaml. Original runner grading and frozen retrospective scoring are separate. Read the frozen source for the rules behind the reported results.
- visits.xlsx exists —
file_exists, required - live formulas —
xlsx_has_formulas, required - no error cells —
xlsx_no_errors, required - merged provider February visits —
xlsx_value_present, required - merged provider quarter visits —
xlsx_value_present, required - second provider no-show count —
xlsx_value_present, required - clinic total visits —
xlsx_value_present, required - no-show rates per provider —
custom, required
Recorded results, desk release
| System | Repetitions | Frozen / 3 | Original / 3 | Median minutes | Mean model estimate (USD) |
|---|---|---|---|---|---|
| Proto / DeepSeek V4.1 Flash | 3 | 3 | 2.5 | 0.0422 | |
| Codex / GPT-5.6 sol | 0 | 0 | 6.7 | 0.6081 |
These are repeated executions on the released fixture, not independent new business cases. Cost is captured model usage priced under the recorded assumptions, excluding infrastructure and human review. Per-attempt check verdicts, raw and frozen verdicts, usage, receipt and artifact hashes for this task are in results/latest/attempts.jsonl under "task": "clinic-visits-summary".
Recorded results, desk comparison, 28 September 2026
| System | Repetitions | Frozen / 3 | Original / 3 | Median minutes |
|---|---|---|---|---|
| Proto / GPT-6 sol | 2 | 2 | 3.4 | |
| Codex / GPT-6 sol | 2 | 2 | 3.5 | |
| Proto / DeepSeek V4.1 Flash | 3 | 3 | 1.7 |
A separate campaign under the same frozen scorer, not pooled with the release. Conditions and data.