Awesomejev Skills

Claude XLSX Skill: Tested Review

The xlsx skill is for opening, fixing and creating spreadsheets. We gave Claude a deliberately messy CSV and checked the finished file against totals we calculated ourselves.

Source
anthropics/skills, folder skills/xlsx
Source commit tested
e4bb5d9
Tested on
2026-11-16 with Claude Code 2.1.325
Runs
One run with the skill, one without

Our verdict

Both spreadsheets were correct: 24 clean rows, a total of 391,003.74 and matching region totals. The skill version added a reconciliation row that checks the two sheets agree. It also needed more than twice the turns and about 70 percent more time in this test.

What it says it does

Use this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file; create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats.

From the skill's own SKILL.md description, source: anthropics/skills at commit e4bb5d9.

What you need before it works

Real tests

Each test ran twice with the same prompt: once with the skill installed, once without. Single run each.

Clean a messy sales CSV and build a chart

Prompt we gave Claude

sales.csv in this folder is messy (blank rows, a repeated header, numbers stored as text like 1,200 and $1,234.56). Turn it into a clean sales.xlsx: a styled header row, Units and Revenue stored as real numbers, a Total row at the bottom that uses a SUM formula, and a bar chart of revenue by region on a second sheet using a formula-based summary. The formulas must have calculated values saved in the file, not just be formulas.
With skillWithout skill
Skill used by ClaudeYes, on its own (xlsx)Not installed
Effort22 agent turns, 105 s, about $0.62 of usage9 agent turns, 61 s, about $0.32 of usage
Data rows24 (correct)24 (correct)
Total revenue391,003.74 (correct)391,003.74 (correct)
Formulas and chartSUM and SUMIFS formulas, bar chart, plus a check rowSUM and SUMIF formulas, bar chart

We opened both files with openpyxl and compared every number with totals we computed from the raw CSV. All matched. We then rendered both in LibreOffice for the screenshots.

The cleaned Sales sheet with styled header, 24 rows and a Total row
With the skill: Sales sheet.
Bar chart of revenue by region with a summary table and a check row
With the skill: summary sheet and chart.
Bar chart of revenue by region with a summary table
Without the skill: summary sheet and chart.

Caveats

How we installed it for the test

We copied the skill folder into .claude/skills/xlsx inside the test project, with the Python libraries available on the PATH.

More tested skills