[
  {
    "type": "title",
    "title": "Presentation Title",
    "subtitle": "Subtitle or author"
  },
  {
    "type": "content",
    "title": "Slide Headline",
    "bullets": ["Point 1", "Point 2", "Point 3"]
  },
  {
    "type": "two_column",
    "title": "Comparison",
    "left": {"header": "Option A", "bullets": ["Pro 1", "Pro 2"]},
    "right": {"header": "Option B", "bullets": ["Pro 1", "Pro 2"]}
  },
  {
    "type": "quote",
    "quote": "Key insight or memorable statement",
    "attribution": "Speaker Name"
  },
  {
    "type": "table",
    "title": "Data Overview",
    "headers": ["Column A", "Column B", "Column C"],
    "rows": [["r1c1", "r1c2", "r1c3"], ["r2c1", "r2c2", "r2c3"]]
  },
  {
    "type": "closing",
    "title": "Thank You",
    "subtitle": "contact@example.com"
  }
]

SLIDE MAP (10 slides, Corporate palette):

  1. [Title] "Q4 Revenue Analysis"
  2. [Content] "Executive Summary" (4 bullets)
  3. [Content] "Revenue by Region" (5 bullets)
  4. [Quote] Key insight from CFO
  5. [Section] "Deep Dive: EMEA"
  6. [Two-Column] EMEA vs APAC comparison
  7. [Content] "Contributing Factors" (4 bullets)
  8. [Table] Quarterly figures
  9. [Content] "Recommendations" (3 bullets)
  10. [Closing] "Questions?"

Approve this structure, or suggest changes?

python3 -c "from pptx import Presentation; print('python-pptx OK')"

pip install python-pptx Pillow

undefined

**Step 3: Run the generation script**

```bash
python3 /path/to/skills/pptx-generator/scripts/generate_pptx.py \
  --slide-map /tmp/slide_map.json \
  --design /tmp/design_config.json \
  --output /absolute/path/to/output.pptx

python3 /path/to/skills/pptx-generator/scripts/validate_structure.py \
  --input /absolute/path/to/output.pptx \
  --slide-map /tmp/slide_map.json

soffice --version 2>/dev/null

python3 /path/to/skills/pptx-generator/scripts/convert_slides.py \
  --input /absolute/path/to/output.pptx \
  --output-dir /tmp/pptx_qa_images/ \
  --dpi 150

You are a visual QA inspector for a PowerPoint presentation. You have ZERO
context about how these slides were generated. Evaluate each slide image
against the QA checklist.

For each slide, check:
1. Text readability (not clipped, not overlapping, adequate size, sufficient contrast)
2. Layout and alignment (consistent margins, aligned elements, visual balance)
3. Color usage (palette consistency, max 3 colors per slide)
4. Content accuracy (title matches expected, content present)
5. Anti-AI violations (accent lines under titles, gradient backgrounds, identical layouts,
   shadows on everything, rounded rectangles everywhere, clip art icons, gradient text)

Return per-slide results:
SLIDE QA RESULT: [PASS | FAIL]
Slides checked: N
Issues found: M

SLIDE 1: PASS | FAIL
  - [Category] Issue description
    FIX: Specific fix instruction
...
OVERALL: PASS | FAIL (N issues on M slides)

QA Iteration 1/3: 2 issues found (1 Blocker, 1 Major)
QA Iteration 2/3: 1 issue found (1 Minor)
QA Iteration 2/3: PASS (0 Blocker, 0 Major)

===============================================================
 PRESENTATION GENERATED
===============================================================

 File: /absolute/path/to/presentation.pptx
 Slides: 10
 Palette: Corporate
 Format: 16:9 widescreen
 Size: 45,230 bytes

 Slide Map:
   1. [Title] "Q4 Revenue Analysis"
   2. [Content] "Executive Summary"
   3. [Content] "Revenue by Region"
   4. [Quote] CFO insight
   5. [Section] "Deep Dive: EMEA"
   6. [Two-Column] EMEA vs APAC
   7. [Content] "Contributing Factors"
   8. [Table] Quarterly figures
   9. [Content] "Recommendations"
  10. [Closing] "Questions?"

 QA Result: PASS (2 iterations, 3 issues fixed)

 Notes:
   - [any remaining minor issues or caveats]

===============================================================

{
  "palette": "minimal",
  "template_path": null
}