Generate, edit, batch-create, inspect, and prepare images for delivery through an OpenAI-compatible image API.
English | 简体中文
What It Does
This portable Agent Skill gives compatible agent clients one workflow for marketing visuals, product cutouts, editorial images, brand assets, interface graphics, game art, and other image deliverables.
Need
Capability
Create or revise images
Text-to-image generation and reference-image editing
Produce controlled variations
JSONL batches with per-task options and limited concurrency
Deliver exact files
PNG resizing, contain/stretch fitting, safe margins, and grid splitting
Check technical requirements
Deterministic qa.v1 checks for size, alpha, edge contact, and optional components
Review presentation contexts
Preview boards at multiple sizes and on multiple backgrounds
Keep credentials local
Ignored auth.json with direct key or environment-variable authentication
The skill uses an OpenAI-compatible image backend. Local post-processing is deterministic and does not call the image API.
Compatibility
The configured base_url must expose these endpoints: