Receivables aging at month end for the bank
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.
Ingrid needs the receivables aging as of 31 August for the bank. Build aging.xlsx with live formulas from the exports and write memo.md she can forward - her note says what goes in both.
Workspace
- credit_memos.csv
- customer_payments.csv
- invoice_register_export_2026-09-08.csv
- note_from_ingrid.txt
Expected deliverables
- aging.xlsx
- memo.md
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 |
|---|---|---|---|
| aging.xlsx exists | file_exists | yes | 0 / 6 |
| live formulas | xlsx_has_formulas | yes | 0 / 6 |
| no error cells | xlsx_no_errors | yes | 1 / 6 |
| total receivables as of 31 August | xlsx_value_present | yes | 2 / 6 |
| 31-60 bucket total | xlsx_value_present | yes | 0 / 6 |
| Pier 9 Fish House balance (September payment not yet received) | xlsx_value_present | yes | 0 / 6 |
| Riverbend Bistro over 90 (unapplied credit on the oldest invoice) | xlsx_value_present | yes | 2 / 6 |
| memo carries total receivables and total past due | text_numbers_present | yes | 0 / 6 |
| memo names the top three past-due customers | custom | yes | 0 / 6 |
Current task declaration (9 checks)
From task.yaml. Original runner grading and frozen retrospective scoring are separate. Read the frozen source for the rules behind the reported results.
- aging.xlsx exists —
file_exists, required - live formulas —
xlsx_has_formulas, required - no error cells —
xlsx_no_errors, required - total receivables as of 31 August —
xlsx_value_present, required - 31-60 bucket total —
xlsx_value_present, required - Pier 9 Fish House balance (September payment not yet received) —
xlsx_value_present, required - Riverbend Bistro over 90 (unapplied credit on the oldest invoice) —
xlsx_value_present, required - memo carries total receivables and total past due —
text_numbers_present, required - memo names the top three past-due customers —
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 | 6.9 | 0.0819 | |
| Codex / GPT-5.6 sol | 1 | 1 | 6.8 | 0.6935 |
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": "ar-aging-report".
Recorded results, desk comparison, 28 September 2026
| System | Repetitions | Frozen / 3 | Original / 3 | Median minutes |
|---|---|---|---|---|
| Proto / GPT-6 sol | 3 | 3 | 4.4 | |
| Codex / GPT-6 sol | 3 | 3 | 4.8 | |
| Proto / DeepSeek V4.1 Flash | 2 | 2 | 2.5 |
A separate campaign under the same frozen scorer, not pooled with the release. Conditions and data.