xlsx

Spreadsheet Processing

Drive everything through Python via run_python_script. The agent has pip_install for any missing dependency. Bundled-friendly libraries: openpyxl (xlsx read/write/formatting), pandas (analysis), csv (stdlib).

Reading Spreadsheets

# read_xlsx.py
import sys
from openpyxl import load_workbook

wb = load_workbook(sys.argv[1], data_only=True)
sheet = wb[sys.argv[2]] if len(sys.argv) > 2 else wb.active
for row in sheet.iter_rows(values_only=True):
    print("\t".join("" if v is None else str(v) for v in row))

For CSV/TSV use the stdlib csv module — no install needed.

Creating Spreadsheets

# create_xlsx.py
from openpyxl import Workbook
from openpyxl.styles import Font

wb = Workbook()
ws = wb.active
ws.title = "Employees"
ws.append(["Name", "Age", "City"])
for cell in ws[1]:
    cell.font = Font(bold=True)
ws.append(["Alice", 30, "Beijing"])
ws.append(["Bob", 25, "Shanghai"])
wb.save("/path/to/output.xlsx")

Data Analysis Workflow

1. Run a script that reads + computes the answer in one go (don't ask the model to mentally aggregate — push the work into Python).
2. For aggregation use pandas; for simple per-row work, plain openpyxl is enough and starts faster.
3. If the user wants a modified version, write rows into a new workbook and save.

Converting Between Formats

• CSV → XLSX: read with csv.reader, write with openpyxl.
• XLSX → CSV: read with openpyxl, write with csv.writer.
• PDF tables → XLSX: use the pdf skill to extract text first, parse rows in Python, then write via openpyxl.

Advanced Operations

Recalculate All Formulas (requires LibreOffice)

When an Excel file has formulas that show stale values or #VALUE!, recalculate them:

python3 scripts/recalc.py input.xlsx output.xlsx

This uses LibreOffice to open the file, recalculate all formulas, and save. Requires soffice CLI.

Python openpyxl (for formatting, formulas, charts)