pptx

This skill enables creation, editing, and analysis of PowerPoint (.pptx) files by treating them as ZIP archives of XML and media resources. It solves the problem of programmatically producing polished presentations and of extracting or manipulating deep presentation features that are not accessible through simple text tools. It provides distinct workflows for reading content, editing existing decks, and generating new presentations from scratch.

For reading, it recommends converting a deck to markdown with markitdown for plain text, and unpacking the archive with a bundled unpack.py script for raw XML access to comments, speaker notes, slide layouts, animations, and complex formatting, documenting the key OOXML file structure (presentation.xml, slides, notesSlides, comments, slideLayouts, slideMasters, theme, media). It explains typography and color extraction from theme files. For creating new decks without a template, it uses an html2pptx workflow that converts styled HTML slides to PowerPoint with accurate positioning via a bundled html2pptx.js library and PptxGenJS, with detailed rules on supported HTML elements, web-safe fonts, shape styling on div elements, and pre-rendering gradients and icons to PNG with Sharp. The package also ships OOXML XSD schemas and Python validation and utility scripts (unpack.py, pack.py, validate.py, replace.py, inventory.py, thumbnail.py).

The target users are people and AI agents who need to generate business or design presentations, extract text and notes from decks, emulate an existing design's typography and palette, or work with layouts, comments, and speaker notes. The scripts run local, read-only or file-transformation operations on presentation files with no network or destructive behavior.