Turn the Q3 leads sheet into a Salesforce lead import
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.
Sales is moving to Salesforce and our Q3 leads need to go in. Please turn leads_master_Q3.xlsx into their import file using the template; the consultant's notes and the users export are in the folder. Save it as sf_leads.csv.
Workspace
- leads_master_Q3.xlsx
- salesforce_admin_notes.txt
- salesforce_users_export.csv
- sf_lead_import_template.csv
Expected deliverables
- sf_leads.csv
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 |
|---|---|---|---|
| Salesforce template columns, exact order | csv_columns | yes | 0 / 6 |
| one row per lead | csv_set_equal | yes | 0 / 6 |
| row count | csv_row_count | yes | 0 / 6 |
| lead source picklist | csv_values_match | yes | 0 / 6 |
| lead status picklist | csv_values_match | yes | 0 / 6 |
| country and state codes | csv_values_match | yes | 0 / 6 |
| lead owner username | csv_values_match | yes | 0 / 6 |
| company placeholder | csv_values_match | yes | 0 / 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.
- Salesforce template columns, exact order —
csv_columns, required - one row per lead —
csv_set_equal, required - row count —
csv_row_count, required - lead source picklist —
csv_values_match, required - lead status picklist —
csv_values_match, required - country and state codes —
csv_values_match, required - lead owner username —
csv_values_match, required - company placeholder —
csv_values_match, required
Recorded results, desk release
| System | Repetitions | Frozen / 3 | Original / 3 | Median minutes | Mean model estimate (USD) |
|---|---|---|---|---|---|
| Proto / DeepSeek V4.1 Flash | 3 | 3 | 1.5 | 0.0192 | |
| Codex / GPT-5.6 sol | 3 | 3 | 2.4 | 0.2646 |
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": "salesforce-lead-import".
Recorded results, desk comparison, 28 September 2026
| System | Repetitions | Frozen / 3 | Original / 3 | Median minutes |
|---|---|---|---|---|
| Proto / GPT-6 sol | 2 | 2 | 1.5 | |
| Codex / GPT-6 sol | 3 | 3 | 1.3 | |
| Proto / DeepSeek V4.1 Flash | 3 | 3 | 1.3 |
A separate campaign under the same frozen scorer, not pooled with the release. Conditions and data.