ppt-read-skill

name	ppt-read-skill
description	Read legacy .ppt and modern .pptx slide decks by converting them to PDF, exporting slide images, and extracting per-slide text. Use when the task is to inspect, summarize, quote, search, or visually review course slides.

PPT Slides Reader

Use this skill for .ppt and .pptx decks, especially legacy .ppt files that need LibreOffice-based rendering.

Quick Start

python3 skills/ppt-read-skill/scripts/read_slides.py "2.1-词法分析概述.ppt"

The command creates an output directory containing:

• deck.pdf: LibreOffice-rendered PDF
• slides/slide-001.jpg: rendered slide images
• slides.md: per-slide text plus image paths
• slides.json: structured metadata and text

Batch Check

python3 skills/ppt-read-skill/scripts/read_slides.py *.ppt *.pptx --out /private/tmp/slides-read

Dependencies

The script uses Codex bundled runtime when present:

• LibreOffice: ~/.cache/codex-runtimes/codex-primary-runtime/dependencies/bin/soffice
• Poppler: ~/.cache/codex-runtimes/codex-primary-runtime/dependencies/bin/pdftoppm
• Python packages: bundled pdfplumber

It also works with system soffice, pdftoppm, and pdfplumber when those tools are installed.

Reading Workflow

1. Convert the deck to PDF with an isolated temporary LibreOffice profile.
2. Render each PDF page to a JPEG slide image.
3. Extract text per page with pdfplumber.
4. Read slides.md for fast text inspection and open slide images for visual checks.

Notes

Legacy .ppt files may duplicate characters in extracted headings after PDF conversion. Use the rendered slide images as the source of truth when exact title text matters.