document-processing — Claude Code skill: PDF, Excel, Word, PowerPoint
A skill for office files: covers PDF, XLSX, DOCX and PPTX, picks the right tool per format and flags the rules that matter.
npx skills add eyadsibai/ltk --skill document-processing -a claude-codeWhat document-processing does
The skill kicks in when a task touches PDF, Excel, Word or PowerPoint, along with XLSX, DOCX, PPTX, spreadsheets and presentations. It maps each job to a tool: pypdf, pdfplumber, reportlab, pytesseract with pdf2image, plus the qpdf and pdftotext command line utilities for PDF; pandas and openpyxl for spreadsheets; pandoc and python-docx for Word; markitdown, python-pptx and pptxgenjs for slides.
Excel gets its own section. The skill is blunt about it: computing a number in Python and writing the result into a cell is wrong, because the value goes stale the moment the data changes. Write the formula instead. It also lists financial-model conventions — blue text for hardcoded inputs, black for formulas, green for links to other sheets, yellow fill for cells that need attention.
Word and PowerPoint come with their file layouts. XLSX, DOCX and PPTX are ZIP archives holding XML, so you can edit the raw parts. For contracts and legal documents the skill recommends tracked changes via the <w:ins> and <w:del> tags rather than overwriting text. Format conversion runs through pandoc and headless LibreOffice. Worth knowing before you install: each task needs its own libraries and utilities, and scenarios that lean on pandoc or LibreOffice will not run without them.
What it helps with
- pull text and tables out of a PDF report
- merge several files into one PDF
- write an Excel formula instead of a static number
- redline a contract with tracked changes
- convert a DOCX to markdown for analysis
Example prompts for the agent
- Parse this PDF and extract every table into a DataFrame
- Put a formula in the xlsx instead of a fixed value so it recalculates
- Edit the docx using tracked changes
How to install document-processing
1Skills CLI (Vercel)
npx skills add eyadsibai/ltk --skill document-processing -a claude-code2Manually: put the files in place
~/.claude/skills/document-processing/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/document-processing/
1Skills CLI (Vercel)
npx skills add eyadsibai/ltk --skill document-processing -a codex2Manually: put the files in place
.agents/skills/document-processing/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add eyadsibai/ltk --skill document-processing -a cursor2Manually: put the files in place
.cursor/skills/document-processing/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install eyadsibai/ltk document-processing2Manually: put the files in place
.github/skills/document-processing/SKILL.md1Skills CLI (Vercel)
npx skills add eyadsibai/ltk --skill document-processing -a gemini-cli2Platform CLI
gemini skills install https://github.com/eyadsibai/ltk --path plugins/ltk-product/skills/document-processingManually: put the files in place
.devin/skills/document-processing/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add eyadsibai/ltk --skill document-processing -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about document-processing
How do I install document-processing in Claude?
Run npx skills add eyadsibai/ltk --skill document-processing -a claude-code. The skill file then lives at ~/.claude/skills/document-processing/SKILL.md.
What does document-processing do?
It tells the agent which tool to use for PDF, Excel, Word and PowerPoint work, and how those formats are structured. Excel formulas and Word tracked changes get their own sections.
Why does the skill insist on Excel formulas over values?
A static number breaks as soon as the source data changes. A formula recalculates on its own, so the skill treats it as the correct approach.
