Documents

xlsx

Try it

Create, edit, and verify Excel workbooks with real formulas, correct number formats, and zero formula errors.

What it does

Handles spreadsheet work end to end with openpyxl for formulas and formatting, pandas for bulk data, and markitdown for a quick read of a sheet. Every output is recalculated through a bundled recalc.py script that runs LibreOffice on a copy and returns JSON listing formula errors and formulas it could not evaluate. The skill also encodes conventions for financial models: input/formula/link color coding, currency and percentage formats, and assumptions kept in their own referenced cells.

When to use it

  • Adding columns, formulas, or formatting to an existing .xlsx or .xlsm file
  • Building a new workbook from scratch or from CSV/TSV data
  • Converting between tabular formats such as CSV and XLSX
  • Auditing a spreadsheet for broken formulas before sending it out

The skill document

XLSX creation, editing, and analysis

TaskApproach
Create or edit with formulas/formattingopenpyxl — see gotchas below
Bulk data in or outpandas (read_excel, to_excel)
Quick look at a sheetmarkitdown file.xlsx — ## SheetName per sheet; reads .xlsm too. No cell coordinates, so don't plan edits from it
Read a model (formulas and values)two load_workbook passes — see gotchas

openpyxl, pandas, and markitdown are preinstalled — do not run pip install first; write the script and import directly. Only if an import fails (or the markitdown command is missing): pip install the missing package.

Script paths below are relative to this skill's directory.

Requirements for every output

  • Professional font (Arial, Times New Roman) throughout, unless the user says otherwise.
  • Zero formula errors. Never ship while recalc.py reports errors_found; the one exception is under not_computed below. If you think an error predates you, prove it: load the original with data_only=True and look at that cell. An error you introduced looks exactly like one you inherited.
  • Use formulas, never hardcoded results. Write sheet['B10'] = '=SUM(B2:B9)', not the Python-computed total. The sheet must recalculate when its inputs change.
  • Follow the user's spec literally. Exact tab names, exact column headers, and the formula they spelled out. A redesign that computes something else fails, however elegant.
  • Document every assumption and hardcoded number where the reader will see it — a cell comment, or an adjacent cell at a table's end. Cite a real source when one exists (Source: Company 10-K, FY2024, Page 45, Revenue Note, [SEC EDGAR URL]); when the number came from the user, say so plainly.
  • A workbook you create for someone to fill in needs a short legend naming which cells to edit, and one example row of realistic values showing the expected format. Never add such a row to a file you were asked to edit.
  • Editing an existing file: match its conventions exactly. They override every guideline here. Find its designated input cells first — a distinct font color, fill, or shading marks them — write only there, and leave every existing formula untouched.

Recalculate (mandatory whenever the file contains formulas)

openpyxl writes formulas as strings with no cached values. Until you recalculate, every formula cell reads back as None to anything reading cached values — pandas, load_workbook(data_only=True), and most previewers.

python scripts/recalc.py output.xlsx [timeout_seconds]   # default 30

LibreOffice computes the formulas on a copy, only their values go into your file, and you get JSON: status (success | errors_found), total_formulas, total_errors, and an error_summary naming up to 100 cells per error type (locations_truncated says how many it withheld — trust total_errors, not the length of the list). Fix what it names and run it again. JSON with an error key instead of a status means nothing was recalculated, and only that case exits non-zero — errors_found exits 0, so never treat a clean exit as a clean workbook.

A green recalc proves your formulas evaluate, not that they are right. An off-by-one range or a reference to the wrong row yields a clean, error-free file with wrong numbers. Write 2–3 formulas first and check they pull the values you expect, before building out a grid.

not_computed lists formulas LibreOffice cannot evaluate, even on success: a function it lacks (_xlfn.XLOOKUP, _xlfn.LET…) or a link to another workbook (='[1]Returns Analysis'!$B$2). They keep what they held: after an openpyxl save, nothing until Excel opens it. For their numbers sooner, read the original with data_only=True; keep the formulas. A sound formula depending on a listed one, directly or not, reports #NAME? too: leave it and tell the user. A cell that tests one (IFERROR(B1,0)) gets LibreOffice's fallback, and a kept value may be stale: say so. After editing sheet XML, delete each changed formula's ``: a mistake that keeps its value is listed, not reported.

Choosing formulas that survive verification

LibreOffice implements fewer functions than Excel, and a formula it cannot evaluate has no value until Excel opens the file.

  • Prefer Excel-2007-era functions — SUMIFS, INDEX, MATCH, IFERROR, SUMPRODUCT — which need no prefix.
  • Six post-2007 functions work, but only with an _xlfn. prefix, because openpyxl writes your formula into the XML verbatim and Excel stores post-2007 names prefixed (its UI hides the prefix): _xlfn.TEXTJOIN, _xlfn.CONCAT, _xlfn.IFS, _xlfn.SWITCH, _xlfn.MAXIFS, _xlfn.MINIFS. Written bare, each yields #NAME?.
  • Never use XLOOKUP, XMATCH, SORT, FILTER, UNIQUE, or SEQUENCE. The runtime's LibreOffice cannot evaluate them, and one that can fills only the top-left cell of their result, since an openpyxl-written file has no spill metadata. Use INDEX/MATCH for lookups, and sort, filter, and de-duplicate in Python before writing the cells.
  • A missing ) is reported as unbalanced brackets. Too few arguments (=ROUND(A1)) is not: LibreOffice returns a value, and Excel may offer to repair the file.

openpyxl gotchas

  • Reading a model takes two loads. data_only=True yields cached values with the formulas gone; the default yields formula strings with no values. One pass cannot give you both.
  • data_only=True is destructive if you save. That workbook has no formulas left, so saving replaces every one with a literal — permanently.
  • data_only=True on a file openpyxl just wrote returns None everywhere — run recalc.py first. (A formula whose result is "" also reads back as None.)
  • Merged cells: write the top-left anchor only. Every other cell in the range is a MergedCell whose .value is read-only.
  • .xlsm loses its macros unless you pass keep_vba=True to load_workbook.
  • A sheet name containing a space must be quoted in a cross-sheet reference: ='Assumptions Inputs'!$B$5. Unquoted, it evaluates to #VALUE!.
  • Saving an upload drops its text boxes, shapes and slicers. First search xl/drawings/ for <xdr:sp (with the space) and look for an xl/slicers/ part; if found, tell the user or edit the sheet XML directly.
  • Nothing mends a malformed workbook: an unclosed quote in a number format ('0.00 "USD'), a defined name starting with =, or duplicate or empty table headers are saved as written, and Excel may offer to repair the file.
  • Partly bold or italic cell text is flattened unless load_workbook(rich_text=True); such a .value is then a CellRichText, not a str.
  • Every new formula needs its own number_format (date, %, currency): nothing infers it, so a date shows as 45337 and 30% as 0.3.
  • On every new chart set chart.roundedCorners = False, .overlay = False on its legend and on each title you gave it (chart.legend, chart.title, chart.x_axis.title…), and, if it has axes (not pie or doughnut), chart.x_axis.delete = False and chart.y_axis.delete = False, or Excel may draw the title or legend over the plot, round the frame, or draw no axes.
  • With a DateAxis, set chart.y_axis.crossAx = 500, or Excel may offer to repair the file.

Financial models

Unless the user says otherwise, or the existing file already does something else.

Color: blue text (0,0,255) for hardcoded inputs and scenario levers · black for formulas · green (0,128,0) for links to another sheet · red (255,0,0) for links to another file · yellow fill (255,255,0) for key assumptions and cells the user should fill in.

Numbers: currency $#,##0, with the unit named in the header (Revenue ($mm)) · zeros render as -, including in percentages ($#,##0;($#,##0);-) · negatives in parentheses · percentages 0.0%, stored as fractions (0.15 renders 15.0%; storing 15 renders 1500.0%) · valuation multiples 0.0x · years as text ("2024", never 2,024).

Structure: every assumption in its own labeled cell, referenced by the formulas that use it (=B5*(1+$B$6), never =B5*1.05) · formulas consistent across every projection period, since a lone edited cell mid-row is the commonest silent error · guard denominators that can be zero.

Dependencies

openpyxl, pandas, markitdown (pip, preinstalled — install only if an import fails or the command is missing) · LibreOffice (soffice, auto-configured for sandboxed environments via scripts/office/soffice.py)

Questions people ask

Does it need Excel installed?
No. Formulas are recalculated by LibreOffice through scripts/recalc.py, which computes them on a copy and writes only the resulting values back into your file.
What happens if a formula uses a function LibreOffice cannot evaluate?
The recalc output lists it under not_computed, for example _xlfn.XLOOKUP or a link to another workbook. Those cells keep whatever they held until Excel opens the file, so the skill reads the original with data_only=True to get their numbers and tells you the values may be stale.
Can it edit a macro-enabled .xlsm file without losing the macros?
Yes, if the workbook is loaded with keep_vba=True. The document also warns that saving an upload drops text boxes, shapes, and slicers, so those are checked for first.
Does a clean recalc mean the numbers are right?
No. A green recalc only proves the formulas evaluate, not that ranges or references are correct. The skill recommends writing two or three formulas first and checking they pull the expected values before building out a full grid.

Related skills

Run Git operations — commits, branches, merges, rebases, conflict resolution, and recovery — with safety rules enforced.

by Iván539 installs31 stars

Turns a rough idea into a finished doc through three guided stages: context gathering, section-by-section drafting, and reader testing.

by Anthropic197.6k stars

pptx

Official

Create, edit, and read .pptx and .potx files, with scripts for text extraction, thumbnails, and validation.

by Anthropic197.6k stars

openai-docs

Official

Get answers about OpenAI APIs and Codex grounded in official developer docs, with citations.

by OpenAI28.8k stars

Get language- and framework-specific security guidance applied to your code, from secure-by-default new code to a prioritized vulnerability report.

by OpenAI28.8k stars

gh-fix-ci

Official

Use when a user asks to debug or fix failing GitHub PR checks that run in GitHub Actions; use `gh` to inspect checks and logs, summarize failure context, draft a fix plan, and implement only after exp

by OpenAI28.8k stars

More from Anthropic

Browse all skills

Turns a rough idea into a finished doc through three guided stages: context gathering, section-by-section drafting, and reader testing.

by Anthropic197.6k stars

docx

Official

Create, edit, read, and verify Word .docx and .dotx files with scripts and XML-level edits.

by Anthropic197.6k stars

pdf

Official

Read, extract, merge, split, rotate, watermark, create, OCR, and password-protect PDF files.

by Anthropic197.6k stars

pptx

Official

Create, edit, and read .pptx and .potx files, with scripts for text extraction, thumbnails, and validation.

by Anthropic197.6k stars

Turns a short art request into a written generative-art philosophy plus a runnable p5.js sketch with seed and parameter controls.

by Anthropic197.6k stars

Draft, test, and iterate on agent skills with qualitative review and quantitative evals.

by Anthropic197.6k stars