Run2 python docx basics

Skill cxcscmu/SkillLearnBench/skills/b2-self-feedback-claude-haiku-4-5/offer-letter-generator/run2_python-docx-basics

[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 run2_python-docx-basics

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

Working with Word documents (DOCX) using python-docx library with proper text replacement strategies.

SKILL.md

3.2 KB, as published. Nobody here has run it

Python-docx for Template Processing

Installation

pip install python-docx

Document Structure

A Word document (.docx) contains:

Document: Root container
Paragraphs: Text blocks with formatting (font, spacing, alignment)
Runs: Individual text segments with consistent formatting within a paragraph
Tables: Structured data with rows and cells

Core Operations

Loading and Saving

from docx import Document

doc = Document('input.docx')
# ... modify document ...
doc.save('output.docx')

Iterating Content

# All paragraphs in document
for para in doc.paragraphs:
    print(para.text)  # Complete text of paragraph

# All tables
for table in doc.tables:
    for row in table.rows:
        for cell in row.cells:
            print(cell.text)  # Text in cell

Text Replacement Pattern

Simple Approach: Paragraph-level

For templates where formatting is not critical, replace at paragraph level:

def replace_in_paragraphs(doc, placeholder, replacement):
    """Replace placeholder in all paragraphs"""
    for para in doc.paragraphs:
        if placeholder in para.text:
            para.text = para.text.replace(placeholder, replacement)

def replace_in_tables(doc, placeholder, replacement):
    """Replace placeholder in all table cells"""
    for table in doc.tables:
        for row in table.rows:
            for cell in row.cells:
                for para in cell.paragraphs:
                    if placeholder in para.text:
                        para.text = para.text.replace(placeholder, replacement)

Note: Assigning to para.text replaces all runs in the paragraph, which removes formatting but is acceptable for template documents with mostly plain text.

Complete Document Replacement

def replace_all(doc, placeholder, replacement):
    """Replace in paragraphs and tables"""
    replace_in_paragraphs(doc, placeholder, replacement)
    replace_in_tables(doc, placeholder, replacement)

Accessing Nested Content

Cells contain Paragraphs

Each table cell contains one or more paragraphs:

for table in doc.tables:
    for row in table.rows:
        for cell in row.cells:
            # Each cell has paragraphs
            for para in cell.paragraphs:
                para.text = 'modified'

Preserving Structure

When iterating to remove elements, use reverse iteration:

# Remove paragraphs in reverse order to avoid index shifting
for idx in range(len(doc.paragraphs) - 1, -1, -1):
    if should_remove(doc.paragraphs[idx]):
        p = doc.paragraphs[idx]._element
        p.getparent().remove(p)

Key Constraints

Text-only approach: Simple para.text assignment removes all run-level formatting
Paragraph boundary: Cannot modify text that spans across paragraphs naturally (it's already split in source)
Conditional markers: Must fit within single paragraph or table cell for regex to work
Placeholder format: Use consistent delimiters like {{KEY}} for easy identification