← powerpoint.md

MiniMax XLSX

by MiniMax-AI · 3.8k (repo)
Skill Popular Repo Excel / XLSX Open Source
Best for: Editing existing Excel files with zero format loss — XML-native means conditional formatting, charts, and styles survive unchanged
Not ideal for: Users who want a simple pip-install tool — the XML workflow requires understanding of OOXML structure for advanced edits

Why XML-Native Matters

5
Operation modes
8
Included tools
0
pip installs required

Most Excel libraries (openpyxl, xlwings) serialize then re-serialize the file — losing conditional formatting, custom styles, and chart details. MiniMax XLSX unzips the .xlsx (which is a ZIP of XML files), edits the XML directly, and rezips. The file is structurally identical to the original except for your changes.

Capability coverage:

Read & analyze
Create from template
XML direct edit
Formula fix
Style audit
Zero format loss
Column add / delete
Row insert / shift
OOXML cheatsheet

5 Operation Modes

Mode When to Use Tools Used
READ Analyze an existing workbook — data, structure, formulas, styles xlsx_reader, style_audit
CREATE Build a new workbook from an XML template xlsx_pack (with template XML)
EDIT Modify an existing file at OOXML level without touching structure xlsx_unpack, xlsx_add_column, xlsx_insert_row, xlsx_shift_rows, xlsx_pack
FIX Repair broken formulas, circular references, or corrupt XML formula_check, xlsx_unpack, xlsx_pack
VALIDATE Audit formulas and styles without making changes formula_check, style_audit

Core Features

📦
xlsx_unpack / xlsx_pack
The core workflow: unzip an .xlsx into its constituent XML files (xl/worksheets/sheet1.xml, xl/styles.xml, etc.), edit them, then repack. This is how Excel itself stores files internally — editing at this level preserves 100% of formatting that higher-level libraries strip out.
🔢
formula_check
Validates every formula in the workbook before and after edits. Detects #REF!, #NAME?, #DIV/0!, circular dependencies, and broken cross-sheet references. Returns a structured report with cell addresses and suggested fixes.
🎨
style_audit
Audits cell styles, font usage, number formats, and conditional formatting rules across all sheets. Useful for enforcing consistent formatting standards or understanding a complex workbook before modifying it.
📖
Built-in OOXML Cheatsheet
Ships with a reference guide to Office Open XML structure — sheet XML schema, cell reference format, style index system, formula encoding. The LLM uses this to write correct XML without hallucinating element names.
📊
xlsx_reader (pandas-backed)
For READ mode: uses pandas for fast structured data analysis. Returns schema, data types, sample rows, pivot summaries, and formula catalog. Faster than XML parsing for data analysis tasks where you don't need to edit.

Install

# No pip install required — pure XML operations git clone https://github.com/MiniMax-AI/skills cp -r skills/minimax-xlsx ~/.claude/skills/minimax-xlsx # Optional: pandas for READ mode (xlsx_reader) pip install pandas openpyxl

The EDIT/CREATE/FIX/VALIDATE modes require no Python packages — they work entirely by manipulating the XML files inside the .xlsx ZIP archive. Only READ mode (xlsx_reader) uses pandas, and that's optional.

Ratings

XML Format Preservation
4.9
Feature Completeness
4.6
Formula Support
4.5
Ease of Use
4.0

Ease is 4.0/5 — no pip install simplifies setup, but understanding XML-level editing has a learning curve for users unfamiliar with OOXML structure.

Signals

Maintainer
MiniMax-AI (official)
Dependencies
None required (pandas optional)
Repo Stars
3.8k (shared with minimax-pptx)
Paired with
minimax-pptx (same repo)

See Also

Anthropic XLSX Skill
Official Anthropic skill, financial modeling standards, zero formula error principle
Excel MCP Server
3.6k stars, MCP protocol, works with any MCP-compatible AI client
Kimi Office Skills
Production-grade XLSX with native charts, data validation, and full OpenXML control