Short Video Cover

This execution layer routes image generation requests to four different API providers—MuleRun Nano Banana 2, APImart GPT Image 2, Atlas Cloud GPT Image 2, and Agnes Image 2.1 Flash—switchable via the `--provider` parameter. It supports both pure text-to-image generation and reference-image-based editing in single or batch modes, designed to be called by other skills rather than used directly.

Edu Subject Infographic transforms any topic into a vertical teaching graphic by automatically identifying the subject, breaking down 3–5 core concepts, assigning a cohesive color scheme, and generating illustrated descriptions for each section. The skill includes dual-layer accuracy verification—checking prompts against subject-specific guidelines before generation, then visually reviewing the rendered PNG for text clarity, formulas, and diagram logic. Batch generation lets you create multiple infographics at once with consistent styling but distinct color palettes.

Agnes AI Skill brings multimodal generation—text, images, and video—to your terminal through the agnes-ai-cli. It wraps Agnes's models (text-to-image, image-to-image, image-to-video, and keyframe video) in a streamlined CLI interface, handling auth, file uploads, and polling automatically. Use it for prototyping, creative iteration, or agent-driven content workflows.

Grok Imagine Image lets you create pictures by calling the grok-imagine-image model through an OpenAI-style `/v1/images/generations` interface. Run the included Python script with your prompt, configure your API endpoint and key, and retrieve generated images saved locally or via public URL.

Analyzes technical articles to pinpoint sections worth illustrating—abstract concepts, processes, comparisons, and system architecture—then generates 16:9 diagrams in your choice of seven visual styles. Supports notebook sketches, professional infographics, cozy handdrawn, tech doodles, cartoon layouts, whiteboard sketches, and executive-tech designs. Keeps long-form content engaging without cluttering every paragraph.

This skill transforms individual subject concepts into two synchronized animated formats rendered by HyperFrames: a narrated teaching video (~90s, 1080p with Chinese voiceover and subtitles) and a silent looping animation (~32s, 1080p for embedding). Both outputs share a single HTML structure, unified storyboard, and thematic color palette across nine subject domains.