Story Character IP Design Board Generator

Problem

In story, game, and animation character design, a single paragraph of description often lacks a stable structure. Direct image generation calls can produce inconsistent views, missing costume details, or style drift. This skill is aimed at turning role settings into a reviewable, reusable standardized character sheet.

Workflow

The skill processes a character concept through a five-step pipeline. It first performs compliance checking and prompt optimization, then extracts structured fields such as character name, art style, appearance, and core persona. It then combines sheet layout rules to generate layout instructions and a final final_prompt, submits the image generation task, and polls the result.

Two calling modes are relevant:
- Simulation mode: without showapi_appkey, it returns a complete preview JSON for quick validation.
- Real call mode: use execute_task.py to submit the task asynchronously and obtain a task_id; then use query_task.py to poll status until the sheet image URL and structured data are returned.

The output usually includes a character sheet with multi-angle views, expression breakdowns, and costume or prop details, plus prompt and asset type information that can be used for further refinement. A reference image can be provided to anchor visual style.

Boundaries

Execution is strictly asynchronous and does not support synchronous return, so the workflow should surface task status while polling. It is more suitable for early design when the character setting is already clear, not for random image generation with no style constraints. When using it, specify personality, appearance, costume, props, and art style clearly to avoid an incomplete sheet.