| name | make-xlsx |
| description | Build an Excel (.xlsx) spreadsheet from rows of data or a TSV/CSV outline. Use when the user asks for a "spreadsheet", "Excel", "xlsx", or to "organize this as a table I can sort in Excel". |
| version | 1.0.0 |
Make an XLSX
Bundled scripts/build_xlsx.py wraps openpyxl (already a Hearth dep
for read_file on .xlsx).
Input format
TAB-SEPARATED. Real tab characters (\t) between fields, NOT commas.
The builder will fall back to comma-detection if you mess it up, but
don't rely on that — write tabs. First row = header.
Name<TAB>Role<TAB>Years
Ada<TAB>Eng<TAB>5
Linus<TAB>Eng<TAB>34
When you write_file, escape tabs literally in the content arg:
"Name\tRole\tYears\nAda\tEng\t5\n..." — NOT comma-separated.
Steps
write_file the TSV to <workspace>/.build/<slug>.tsv (scratch).run_command:
python <hearth>/skills/make-xlsx/scripts/build_xlsx.py \
--tsv <workspace>/.build/<slug>.tsv \
--out <workspace>/XLSX/<slug>.xlsx \
--sheet "<sheet name>"
- Script prints the output path on stdout.
- Open it for them.
run_command('powershell -Command "Invoke-Item ''<full path>''"') so Excel pops it open.
- Delete the TSV scratch via
delete_path on <workspace>/.build/<slug>.tsv after the .xlsx exists.
Style defaults
- Header row bold + freeze pane
- Auto-fit column widths (first 30 rows sampled)
- Currency / number / date columns NOT auto-formatted — the model can
request a follow-up format pass if needed
Don't
- Don't try to add charts here. If the user wants a chart, generate a
PNG with matplotlib and reference it from a doc, or use the chart
options in Excel after the file lands.
- Don't write 50K rows of made-up data. Confirm the row count with the
user if it's over 200.
