PPT Parse Analyzer Skill

A two-stage PowerPoint toolkit: lossless PPTX extraction first, template intelligence second.

Overview

This repository provides an open-source workflow for PowerPoint template analysis. It first unpacks a .pptx file into a lossless OpenXML-oriented workspace with structured indexes, then analyzes that parsed result to generate reusable design insights such as theme colors, typography, layout systems, placeholders, media usage, and slide patterns.

Repository Structure

• ppt-parser/: Parses PPTX archives into a deterministic analysis workspace.
• ppt-template-analyzer/: Analyzes parsed output and generates template reports.

Workflow

1. Run ppt-parser on a .pptx file.
2. Review parser outputs (manifest, summary, raw/ extraction).
3. Run ppt-template-analyzer on the parsed output directory.
4. Review generated Markdown and JSON reports.

Quick Start

Run commands from each skill directory.

1) Parse PPTX

node scripts/parse_pptx.js <path-to-pptx-file>

Parser output directory naming convention:

<PPT_NAME>_<YYYYMMDD>/

2) Analyze Template

node scripts/analyze_template.js <parsed-pptx-folder> [report-output.md|analysis-output.json]

If output names are omitted, analyzer defaults to:

<PPT_NAME>_<YYYYMMDD>_template_report.md
<PPT_NAME>_<YYYYMMDD>_template_report.json

Output Contract

The parsed directory should include:

<PPT_NAME>_<YYYYMMDD>/
  <PPT_NAME>_<YYYYMMDD>_manifest.json
  <PPT_NAME>_<YYYYMMDD>_summary.json
  <PPT_NAME>_<YYYYMMDD>_summary.txt
  raw/
    [Content_Types].xml
    _rels/.rels
    ppt/
      presentation.xml
      _rels/presentation.xml.rels
      slides/
      slides/_rels/
      slideLayouts/
      slideLayouts/_rels/
      slideMasters/
      slideMasters/_rels/
      theme/
      charts/
      diagrams/
      notesSlides/
      media/
  images/
  media/

Key guarantees:

• raw/ preserves archive internal paths exactly, including _rels.
• manifest.json is machine-focused indexing metadata.
• summary.json is lightweight analysis handoff data.
• images/ and media/ are convenience copies; raw/ppt/media/ is source of truth.

Scope and Boundaries

• The parser focuses on extraction and indexing, not design interpretation.
• The analyzer provides structured template heuristics, not final visual judgment.
• XML relationships (.rels) should be treated as authoritative linkage data.

License

Add your preferred open-source license (for example, MIT) before publishing.