Pitfall Guide Image Generation

The Problem

Creating pitfall-guide images for Xiaohongshu posts, Moments shares, or video covers usually requires turning a topic into key points, risks, and warnings, then styling a clean card with illustration and concise text. Manually refining copy, choosing an illustration style, and controlling whitespace can be repetitive. This skill compresses the input into one required question, such as "eating street food at a night market," and generates a clean flat-illustration pitfall-guide image, reducing repeated content-asset work.

How It Works

The skill works through ShowAPI's image-generation API using asynchronous task submission + polling.
- Compliance check and prompt optimization: validate the topic and convert the user's description into drawing instructions.
- Async submission: scripts/execute_task.py submits the task and returns only a task_id without blocking.
- Polling: scripts/query_task.py queries task status until the final image URL is available; each poll response should be surfaced to the user.
- Presentation: the result can be rendered as a simple HTML page that combines the image and core pitfall points.

The output style is Morandi low-saturation, whitespace card layout, and concise text, supporting 3:4, 16:9, and 1:1, with 3:4 as the default. It fits Xiaohongshu posts, posters, Moments shares, and video covers.

Boundaries

Without showapi_appkey, the system runs in simulation mode and returns a one-time preview result without execute_task.py or query_task.py. That output is only for previewing the effect, not real generation. Real calls require a ShowAPI account and AppKey. The skill is asynchronous only and does not support synchronous generation; question must clearly describe the target item, otherwise the generated emphasis may drift.