xlsx¶
What it does¶
Reads, creates, and edits spreadsheet files — .xlsx, .xlsm, .csv, .tsv. Handles things like:
- Adding columns, computing formulas, formatting cells
- Cleaning up messy data (malformed rows, misplaced headers, junk values)
- Creating charts and pivot summaries
- Converting between tabular file formats
For data inside Google Sheets directly, use the gsheets MCP instead. This skill is for standalone files that live on disk.
When to use it¶
- A client sent you a spreadsheet that needs cleanup before you can use it
- You need to produce a deliverable that has to be an
.xlsxfile (not Google Sheets) - Tabular data lives in a
.csvand needs to become a.xlsxwith formulas / formatting - You need to convert between formats (CSV → XLSX, TSV → CSV, etc.)
How to invoke¶
You: clean up the spreadsheet at ~/Downloads/client-data.xlsx — the
headers are on row 3, there are junk rows above them, and column F
has dates in two different formats.
Claude: [reads the file, cleans it up, saves a cleaned version]
Trigger phrases:
- "clean this spreadsheet"
- "add a column to [file]"
- "compute [formula] in this xlsx"
- "convert this .csv to xlsx with formatting"
What you get back¶
- A cleaned, modified, or new spreadsheet saved to disk
- A summary of what was changed (rows cleaned, columns added, formulas applied)
- Optionally a backup of the original (always asked first before overwriting)
Example session¶
You: clean up ~/Downloads/raw-creator-list.csv. Headers should be on
row 1. Add a column for "Followers (numeric)" that parses the
Followers column where it's currently strings like "287K" and
"1.2M". Save as xlsx with the numeric column right-aligned.
Claude: Reading raw-creator-list.csv...
Detected: 247 rows, 8 columns, headers on row 1 (already correct).
Adding "Followers (numeric)" column:
- "287K" → 287000
- "1.2M" → 1200000
- "456" → 456
- 3 rows with unparseable values flagged with #N/A
Saving as raw-creator-list-cleaned.xlsx with the new column
right-aligned and number-formatted.
Done. File saved.
Limitations¶
- Doesn't work with Google Sheets (use gsheets MCP for that)
- Doesn't preserve macros / VBA code if you save and reopen
- Complex pivot tables may not round-trip cleanly through edit-and-save
- For huge files (100K+ rows), processing slows down significantly