File Conversion Skill

Convert documents to clean Markdown. Small files output directly; large files return a document map with extraction options.

When to Use

• User shares or references a .docx, .pdf, .xlsx, .pptx, or .csv file
• User asks to "read", "analyze", "summarize", or "extract" content from a document
• You encounter a binary file format you cannot read directly

When NOT to Use

• Plain text files (.txt, .md, .json, .yaml) — read directly
• Source code files — read directly
• Image files — use vision capabilities instead

Workflow

1. Run shuck

python "{{PLUGIN_DIR}}/shuck.py" "<path>"

• Small file → full Markdown output (use directly)
• Large file → document map with sections, token counts, and suggested next steps

2. Follow the map (large files only)

The map suggests commands. Pick based on the task:

# Get specific sections
python "{{PLUGIN_DIR}}/shuck.py" "<path>" --sections s1,s3

# Search for keywords
python "{{PLUGIN_DIR}}/shuck.py" "<path>" --grep "revenue"

# Tables only
python "{{PLUGIN_DIR}}/shuck.py" "<path>" --tables-only

# Fit within a token budget
python "{{PLUGIN_DIR}}/shuck.py" "<path>" --budget 4000

# Force full output
python "{{PLUGIN_DIR}}/shuck.py" "<path>" --all

# Excel/CSV: schema or sample
python "{{PLUGIN_DIR}}/shuck.py" "<path>" --schema-only
python "{{PLUGIN_DIR}}/shuck.py" "<path>" --sample 5

Installing Dependencies

pip install python-docx pdfplumber openpyxl python-pptx