PPTX Presentation Toolkit

Problem addressed

When producing .pptx files, the harder issue is rarely finding a template; it is reliably extracting text, analyzing layouts, reusing masters, reserving chart areas, inheriting fonts and colors, and validating visual output. Editing XML directly can corrupt OOXML structure, while rendering HTML into PowerPoint requires precise positioning. This skill treats PPTX as a ZIP container of XML and resources, oriented toward programmatic creation, editing, and analysis.

Workflow and core capabilities

• Reading and analysis: extract body text with markitdown, then inspect ppt/theme/theme1.xml, ppt/slides/slideN.xml, notesSlides, and comments for fonts, colors, speaker notes, and comments.
• Creation from scratch: define the design intent, generate HTML per slide, use placeholders for chart or table regions, and combine html2pptx, PptxGenJS, and Sharp to produce precise layouts; validate thumbnails for truncation, overlap, and contrast.
• Template reuse: build template-inventory.md and a text inventory, use rearrange.py to duplicate and reorder slides, use inventory.py to extract shape metadata, and use replace.py to validate and write replacement paragraphs.
• Editing existing files: unpack, modify slide{N}.xml, run validate.py, then repack; suitable for speaker notes, layouts, animations, and other XML-level details.

Boundaries and notes

The skill favors reproducible file-level operations for engineers automating PPTX generation or modification. For complex themes, embedded media, animations, or strict brand standards, preserve the original XML structure and perform visual validation. Before replacing text, confirm actual shapes, font sizes, and placeholder fit, and avoid forcing content into mismatched multi-column layouts.