odp-slide-generator

​

When to use this skill

​

Format Rules

1. Create a markdown file with the following exact structure:
   • # [Main Slide Title] (Must be Top Level)
   • ## [Slide Subtitle] (Optional Subtitle)
   • ### [Column Label] (e.g. Stage 01: Source)
   • **[Box Header Text]** (e.g. Direct Extraction)
   • - [Bullet point 1]
   • - [Bullet point 2]
   • > "[Footer Quote Text]" (Optional footer quote block)

​

Example Markdown:

# Achieving Zero-Downtime Cutover
## Near-Real-Time Delta Sync (CDC) Architecture

### Stage 01: Source
**Direct Extraction**
- Native connection to PostgreSQL Write-Ahead Logs (WAL).
- Row-level mutation tracking without database triggers.

### Stage 02: Transport
**Streaming Fabric**
- Apache Kafka (KRaft Mode) for high-throughput event bus.
- Debezium Connect for automated change capture.

### Stage 03: Target
**Instant Mirroring**
- Near-instant transaction mirroring to Percona target.
- Continuous data validation during sync phase.

> "By streaming row-level mutations in real-time, we compress the final production cutover window."

​

Execution

1. Write the target markdown file (e.g., docs/slide_data.md).
2. Run the compiler using uv:
   uv run --with python-pptx python tools/compile_3col_slides.py docs/slide_data.md docs/my_presentation.pptx
   
3. The resulting .pptx file is 100% natively compatible with LibreOffice Impress and will retain all shape geometries, exact RGB colors, and custom fonts.