Docx template fill

Skill cxcscmu/SkillLearnBench/skills/b4-skill-creator-claude-opus-4-6/offer-letter-generator/docx-template-fill

[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 docx-template-fill

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

Fill Word document (.docx) templates that use placeholder syntax like {{PLACEHOLDER}}. Handles split placeholders (where Word splits a placeholder across multiple XML runs), conditional sections ({{IF_X}}...{{END_IF_X}}), and preserves formatting. Use this skill whenever the user asks to fill a Word template, generate a document from a template, or replace placeholders in a .docx file.

SKILL.md

3.5 KB, as published. Nobody here has run it

DOCX Template Fill

Core Challenge: Split Placeholders

Word internally splits text into "runs" based on formatting changes, spell-check boundaries, or editing history. A placeholder like {{COMPANY_NAME}} may become three runs: {{COMP + ANY_ + NAME}}. Naive per-run replacement will fail.

Strategy

Merge-then-replace at the run level: For each paragraph (and header/footer/table cell paragraph), concatenate all run texts, perform regex replacements on the merged string, then redistribute text back to runs — putting all text in the first run and clearing the rest. This preserves the first run's formatting.
Process all text containers: Paragraphs live in the document body, but also in:
- Table cells (iterate doc.tables[*].rows[*].cells[*].paragraphs)
- Section headers (doc.sections[*].header.paragraphs)
- Section footers (doc.sections[*].footer.paragraphs)
Conditional sections: For {{IF_KEY}}...content...{{END_IF_KEY}}:
- If condition is true (value is "Yes"), remove only the markers, keep content
- If condition is false, remove markers and all content between them
- Conditionals may span a single paragraph or the markers may be in separate runs

Implementation Pattern

import re, json
from docx import Document

def replace_in_paragraph(paragraph, replacements):
    """Replace placeholders in a paragraph, handling split runs."""
    full_text = "".join(run.text for run in paragraph.runs)
    if not full_text:
        return
    new_text = full_text
    for key, value in replacements.items():
        new_text = new_text.replace("{{" + key + "}}", value)
    if new_text != full_text:
        # Put all text in first run, clear the rest (preserves first run's formatting)
        for i, run in enumerate(paragraph.runs):
            if i == 0:
                run.text = new_text
            else:
                run.text = ""

def handle_conditional(paragraph, key, keep):
    """Handle {{IF_key}}...{{END_IF_key}} within a single paragraph."""
    full_text = "".join(run.text for run in paragraph.runs)
    pattern = r"\{\{IF_" + key + r"\}\}(.*?)\{\{END_IF_" + key + r"\}\}"
    if re.search(pattern, full_text):
        if keep:
            new_text = re.sub(pattern, r"\1", full_text)
        else:
            new_text = re.sub(pattern, "", full_text)
        for i, run in enumerate(paragraph.runs):
            run.text = new_text if i == 0 else ""

def process_all_paragraphs(doc, func, *args):
    """Apply a function to every paragraph in the document."""
    for p in doc.paragraphs:
        func(p, *args)
    for table in doc.tables:
        for row in table.rows:
            for cell in row.cells:
                for p in cell.paragraphs:
                    func(p, *args)
    for section in doc.sections:
        for p in section.header.paragraphs:
            func(p, *args)
        for p in section.footer.paragraphs:
            func(p, *args)

Workflow

Load data (JSON) and template (docx)
Handle conditionals first (before placeholder replacement)
Replace all {{KEY}} placeholders with values from data
Save the filled document