baoyu-slide-deck

The baoyu-slide-deck skill transforms raw text content into high-quality, professional slide deck images optimized for digital consumption and social media sharing. Unlike traditional presentation software designed for live speaking, this tool focuses on self-explanatory, readable, and visually consistent slides that emphasize logical scroll flow and information density. It is an essential utility for content creators, product managers, and educators who need to repurpose long-form articles or notes into engaging visual summaries.

• Automated pipeline: Generates slide outlines based on user text, applies professional style instructions, and renders individual slides as images.

• Rich style system: Includes 17 design presets such as blueprint, corporate, notion, and minimal, covering diverse use cases from technical system design to lifestyle tutorials.

• Flexible configuration: Supports command-line options for audience targeting (beginners to executives), output language settings, slide counts, and reference image styles.

• Reproducibility: Enforces a strict workflow where final prompts are saved as individual files, enabling users to swap image backends or iterate on specific slides without full re-generation.

• Formatting and export: Provides specialized scripts for merging generated slides into unified PowerPoint (.pptx) or PDF documents for seamless distribution.

• Use cases: Converting research papers into executive summaries, creating product update decks for SaaS, or turning technical blog posts into carousel-style image guides.

• Input requirements: Accepts raw text or markdown files as input. The tool uses a clear priority queue for backend image generation, allowing users to override default settings for specific needs.

• Operational notes: Always confirms generation before starting, though this can be overridden with flags like --yes. The tool supports batching of user inputs to minimize friction during setup. Users should set their preferred image backends in EXTEND.md or via current request flags. The output is strictly optimized for 8-25 slide sequences, with a hard limit of 30 slides to ensure quality and coherence.