Python docx

Skill cxcscmu/SkillLearnBench/skills/b1-one-shot-claude-sonnet-4-6/offer-letter-generator/python-docx

[COLM'26] SkillLearnBench is the first benchmark for evaluating continual learning methods that automatically generate agent skills.

Install

npx -y skills add cxcscmu/SkillLearnBench --skill python-docx

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

What its author says it does

Copied from the file, not written here

Core python-docx usage for reading, modifying, and saving Word documents including paragraphs, runs, tables, headers, and footers.

SKILL.md

2.7 KB, as published. Nobody here has run it

python-docx Core Usage

Installation

pip install python-docx

Basic Document Operations

from docx import Document

# Open existing document
doc = Document('template.docx')

# Access paragraphs
for para in doc.paragraphs:
    print(para.text)           # Full paragraph text
    for run in para.runs:      # Individual styled runs
        print(run.text)

# Save
doc.save('output.docx')

Document Structure

A .docx file has:

doc.paragraphs — body paragraphs (does NOT include headers/footers)
doc.tables — top-level tables in body
doc.sections — page sections (each has .header and .footer)

Paragraphs and Runs

A Paragraph is a block of text. It contains multiple Run objects, each with their own formatting (bold, italic, font size, etc.).

Critical: Word often splits a single logical text segment across multiple runs (due to spell-check marks, formatting, or XML internals). Always read para.text for the full text — never rely on individual runs.

para.text          # Full concatenated text of all runs
para.runs          # List of Run objects
run.text           # Text of this run
run.bold           # Bold formatting
run.italic         # Italic formatting
run.font.size      # Font size (in EMUs; divide by 12700 for pt)
run.font.name      # Font name

Tables

for table in doc.tables:
    for row in table.rows:
        for cell in row.cells:
            for para in cell.paragraphs:
                print(para.text)
            for nested_table in cell.tables:  # recurse for nesting
                pass

Headers and Footers

for section in doc.sections:
    header = section.header
    footer = section.footer
    for para in header.paragraphs:
        print(para.text)
    for para in footer.paragraphs:
        print(para.text)

Modifying Text While Preserving Formatting

When replacing text, preserve the first run's formatting and clear subsequent runs:

def set_paragraph_text(para, new_text):
    """Replace all run text in a paragraph, keeping first run's formatting."""
    if not para.runs:
        return
    para.runs[0].text = new_text
    for run in para.runs[1:]:
        run.text = ''

Common Pitfalls

doc.paragraphs does NOT include paragraphs inside table cells
doc.paragraphs does NOT include header/footer paragraphs
Split runs are the norm, not the exception — always work at para.text level
Clearing all runs and rebuilding loses per-run formatting (bold/italic on parts of text)