Article & Presentation Visuals
Turn judgments in a finished narrative into placed images or continuous presentation frames.
Open for discussion · MIT v0.8.1
See delivery and example
Decorative images leave readers to untangle mechanisms, comparisons, and boundaries on their own.
Creators, teams, consultants, and educators who explain complex ideas through articles, reports, talks, proposals, or courses.
Start with existing content. Deliver images with clear placements.
You provide
- Local Markdown; Word / Docs exports or authorized Feishu documents first become a Markdown working copy
- Optional existing assets you may use
The Skill delivers
- Article mode: cover and explanatory images at approved reading anchors
- Presentation mode: cover, optional agenda, ordered frames, and closing
- Page manifest recording source support, prompts, images, placement, and QA
Review the plan before generating images.
See all visual grammars and the workflow ↗- 01Locate the comprehension obstacle and source support
- 02Confirm the visual plan and selected grammar
- 03Generate and finalize each page, then record placement
One source passage, one explanatory image
The public demo article explains why a shoe can feel dry outside while moisture remains inside. The image maps that claim onto the shoe’s structure.
Image from the public smkt-offer demo article. It shows a produced sample, not full-workflow validation on every host. Read the source article (Chinese) ↗
Ask your Agent to install this open-source Skill.
Please install this Skill: https://github.com/Lone3m-tech/smkt-article-visual
Codex completed the full workflow; WorkBuddy runs with visual detail loss; Doubao has not passed the full workflow.
Requires local file access, image generation, Python 3.10+, and Pillow. It does not write prose or deliver editable PPTX files; source assets must be authorized.