FY2026 fundraising figures for the board
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.
Nadia needs this year's fundraising figures for the board packet - totals against last year, the top ten donors and donor retention. Put them in annual_figures.xlsx with live formulas and write a cover memo as memo.md. Her email and the gift policy are in the folder.
Workspace
- email_from_treasurer.txt
- gift_reporting_policy.txt
- gifts_export_2024-01-01_to_2026-06-30.csv
Expected deliverables
- annual_figures.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 |
|---|---|---|---|
| annual_figures.xlsx exists | file_exists | yes | 0 / 6 |
| live formulas | xlsx_has_formulas | yes | 0 / 6 |
| no error cells | xlsx_no_errors | yes | 2 / 6 |
| FY2026 net cash raised | xlsx_value_present | yes | 0 / 6 |
| FY2025 net cash raised | xlsx_value_present | yes | 1 / 6 |
| top donor FY2026 total | xlsx_value_present | yes | 0 / 6 |
| retained donors | xlsx_value_present | yes | 0 / 6 |
| donor retention rate | custom | yes | 0 / 6 |
| memo carries both years' totals | text_numbers_present | yes | 0 / 6 |
| memo says the foundation gift was refunded | text_sentence_matches | yes | 1 / 6 |
Current task declaration (10 checks)
From task.yaml. Original runner grading and frozen retrospective scoring are separate. Read the frozen source for the rules behind the reported results.
- annual_figures.xlsx exists —
file_exists, required - live formulas —
xlsx_has_formulas, required - no error cells —
xlsx_no_errors, required - FY2026 net cash raised —
xlsx_value_present, required - FY2025 net cash raised —
xlsx_value_present, required - top donor FY2026 total —
xlsx_value_present, required - retained donors —
xlsx_value_present, required - donor retention rate —
custom, required - memo carries both years' totals —
text_numbers_present, required - memo says the foundation gift was refunded —
text_sentence_matches, required
Recorded results, desk release
| System | Repetitions | Frozen / 3 | Original / 3 | Median minutes | Mean model estimate (USD) |
|---|---|---|---|---|---|
| Proto / DeepSeek V4.1 Flash | 1 | 1 | 13.0 | 0.1285 | |
| Codex / GPT-5.6 sol | 1 | 1 | 9.2 | 0.8874 |
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": "donor-annual-figures".
Recorded results, desk comparison, 28 September 2026
| System | Repetitions | Frozen / 3 | Original / 3 | Median minutes |
|---|---|---|---|---|
| Proto / GPT-6 sol | 0 | 0 | 3.0 | |
| Codex / GPT-6 sol | 1 | 1 | 4.6 | |
| Proto / DeepSeek V4.1 Flash | 2 | 2 | 3.4 |
A separate campaign under the same frozen scorer, not pooled with the release. Conditions and data.