What problem does it solve? Generating images through multiple AI provider APIs requires juggling different authentication schemes, model IDs, size rules, and reference-image capabilities. This Skill unifies text-to-image and reference-based image generation behind one CLI with automatic provider detection, retries, and batch parallelism. ## Core Features & Use Cases - Multi-Provider Generation: Supports OpenAI GPT Image 2, Azure OpenAI, Google, OpenRouter, DashScope, Z.AI, MiniMax, Jimeng, Seedream, Replicate, and a Codex CLI provider, with automatic provider selection based on available API keys. - Reference-Image Identity Preservation: Accepts reference images on ref-capable providers and provides prompting guidance to preserve subject identity across generations. - Batch Parallel Generation: Runs multi-image jobs from a JSON batch file with configurable worker counts, per-provider rate limits, and up to 3 retries per image. - Use Case: A content team saves finalized prompts from an article outline, builds a batch.json with scripts/build-batch.ts, and generates all illustrations in parallel through Replicate with consistent aspect ratio and quality settings. ## Quick Start Ask the agent to generate an image of your subject with your chosen aspect ratio and save it to a specific file, for example: generate a 16:9 image of a cat and save it as cat.png.
