article-batch-illustration_skill

Overview

This skill analyzes a Markdown article and batch-generates AI illustrations for each H2/H3 section. It creates structured image prompts, calls the Gemini image API to produce images in one of two selectable styles, saves outputs to an Obsidian image folder, and inserts wikilink references into the document. The workflow focuses on consistent, professional logic/concept diagrams tailored to each paragraph.

How this skill works

The tool parses the article, splits by H2/H3 headings, and extracts each section's title, one-line core idea, and key concepts. It constructs a global STYLE INSTRUCTIONS block and per-section English prompts that include Chinese labels and a mandatory instruction to keep image text primarily Simplified Chinese. After user confirmation, it calls a generation script to create images, saves them under a dated Obsidian directory, and inserts ![[...]] links beneath each heading.

When to use it

• When you want consistent concept diagrams for a long Markdown article
• When preparing teaching materials, documentation, or blog posts needing visual summaries
• When you need many schematic images with a unified style quickly
• When migrating visuals into an Obsidian vault with automatic insertion
• When you want Chinese-labelled diagrams with technical English terms preserved

Best practices

• Structure the article with clear H2/H3 headings for reliable splitting
• Review and edit the generated prompt list before batch generation
• Choose the global style early; override per-image only when needed
• Keep each section focused—one main idea per heading yields clearer diagrams
• Verify saved filenames and Obsidian path to avoid broken links

Example use cases

• Generate 10 concept diagrams for a technical blog post and insert them into the Markdown file
• Convert a product spec document into a visual overview set using blueprint style for architecture slides
• Create teaching visuals for course notes with hand-drawn style and Chinese labels
• Regenerate a single figure after editing its prompt without reprocessing the whole article
• Batch process internal docs and save dated image directories for versioning