How do I install the Spreadsheets skill?

Run npx skillmds@latest add openinterpreter/spreadsheets in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Spreadsheets safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Spreadsheets?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Spreadsheets free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Spreadsheets?

openinterpreter (@openinterpreter) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Spreadsheets by openinterpreter — Agent Skill

Name: Spreadsheets
Author: openinterpreter

Spreadsheets

Create, inspect, analyze, and edit spreadsheets (`.xlsx`, `.xlsm`, `.csv`, `.tsv`) with local code execution. Use for formulas, formatting, charts, validation, financial models, and spreadsheet deliverables.

openinterpreter 138d32f 3 files · 13.2 KB Updated Sep 17, 2026

File contents

Spreadsheet workflow

This is workflow guidance, not a callable tool. Use the shell or code-execution tool exposed by the active OIX harness. Do not look for a tool named spreadsheets.

Workflow

Inspect workbook sheet names, used ranges, formulas, styles, merged cells, tables, and charts before editing.
Use openpyxl for .xlsx/.xlsm creation and edits. Load macro-enabled workbooks with keep_vba=True when macros must survive. Use pandas only for analysis or reshaping, not as the final formatting layer.
Build substantial workbooks in one cohesive script: create inputs and source sheets first, then formulas, formatting, validation, tables, charts, and print settings.
Keep business logic in auditable formulas and visible assumption cells. Do not hardcode derived values.
Reopen the saved workbook and verify sheet names, key values, formulas, defined names, and expected structures. Scan formulas for obvious broken references.
Formula libraries do not calculate cached results. When recalculation or visual review is required, use an installed permissive office renderer in headless mode, then reopen and inspect the recalculated file. State the limitation if no renderer is available.
Render or export every important sheet for a visual pass when possible.
If the workbook is open in Interpreter, call interpreter-app tools builtin-interpreter interpreter_refresh_file --json '{"path":"/absolute/path.xlsx"}' once after the final write.

Dependencies

uv pip install openpyxl pandas

Install only what the task needs. Use an existing Python installation when available, and request approval for exact system-level installation commands.

Quality bar

Preserve existing workbook structure and style unless redesign is requested.
Use real formulas, tables, filters, freeze panes, validation, conditional formats, and charts when they improve the requested result.
Avoid formula errors, clipped labels, unreadable widths, and default blank sheets.
For financial or operational models, include visible assumptions and an independent check of sample outputs.
Keep scratch CSV, scripts, and previews outside the final deliverable folder.

---name: spreadsheetsdescription: Create, inspect, analyze, and edit spreadsheets (`.xlsx`, `.xlsm`, `.csv`, `.tsv`) with local code execution. Use for formulas, formatting, charts, validation, financial models, and spreadsheet deliverables.---# Spreadsheet workflowThis is workflow guidance, not a callable tool. Use the shell or code-executiontool exposed by the active OIX harness. Do not look for a tool named`spreadsheets`.## Workflow1. Inspect workbook sheet names, used ranges, formulas, styles, merged cells, tables, and charts before editing.2. Use `openpyxl` for `.xlsx`/`.xlsm` creation and edits. Load macro-enabled workbooks with `keep_vba=True` when macros must survive. Use `pandas` only for analysis or reshaping, not as the final formatting layer.3. Build substantial workbooks in one cohesive script: create inputs and source sheets first, then formulas, formatting, validation, tables, charts, and print settings.4. Keep business logic in auditable formulas and visible assumption cells. Do not hardcode derived values.5. Reopen the saved workbook and verify sheet names, key values, formulas, defined names, and expected structures. Scan formulas for obvious broken references.6. Formula libraries do not calculate cached results. When recalculation or visual review is required, use an installed permissive office renderer in headless mode, then reopen and inspect the recalculated file. State the limitation if no renderer is available.7. Render or export every important sheet for a visual pass when possible.8. If the workbook is open in Interpreter, call `interpreter-app tools builtin-interpreter interpreter_refresh_file --json '{"path":"/absolute/path.xlsx"}'` once after the final write.## Dependencies```shuv pip install openpyxl pandas```Install only what the task needs. Use an existing Python installation whenavailable, and request approval for exact system-level installation commands.## Quality bar- Preserve existing workbook structure and style unless redesign is requested.- Use real formulas, tables, filters, freeze panes, validation, conditional formats, and charts when they improve the requested result.- Avoid formula errors, clipped labels, unreadable widths, and default blank sheets.- For financial or operational models, include visible assumptions and an independent check of sample outputs.- Keep scratch CSV, scripts, and previews outside the final deliverable folder.

Frequently asked questions