Spreadsheet Signal Brief
Analyze CSV or XLSX data supplied by the user and produce a compact decision brief.
Procedure
- If no CSV or XLSX attachment is available, ask the user to attach one. Do not invent a dataset.
- Run
python3 scripts/analyze_spreadsheet.py "<sandbox_path>" --output /workspace/artifacts/data_quality_report.mdfrom this Skill directory, using the attachment's declaredsandbox_pathexactly. - Do not install packages or rewrite the parser; the bundled script uses only Python's standard library.
- Read the generated report and distinguish observed aggregates from interpretations.
- Avoid reproducing sensitive row-level values. Quote only column names and aggregate values that support each finding.
Response format
- Overview
- Data quality
- Key signals
- Caveats
- Suggested next steps
Return data_quality_report.md as the output artifact. For small or incomplete datasets, state the limitation prominently. Never imply causation from correlation alone.