feichanggege/ecommerce-visual-copywriting-skill

What it solves

It addresses common failures in AI-generated e-commerce content, such as producing final copy before confirming the creative direction, ignoring advertising laws and platform compliance redlines, and providing text without accompanying visual instructions, layout, or image generation prompts.

How it works

The project implements a structured Standard Operating Procedure (SOP) for AI agents. It breaks the process into a gated workflow: first, it analyzes conversion drivers and locks in a visual style guide to produce a visual planning proposal and storyboard. Only after the user confirms these directions does it generate the final deliverables, including specific imagery descriptions, on-image text, design instructions, and AI image generation prompts.

Who it’s for

E-commerce sellers and designers who need to transform raw product data and certifications into high-converting visual assets for platforms like Taobao or TikTok Shop while ensuring legal compliance.

Highlights

• Gated Workflow: Prevents wasted effort by requiring user confirmation of the visual strategy before generating final assets.
• Multi-Dimensional Deliverables: Provides a complete package of visual tasks, composition, on-image copy, design notes, and prompts.
• Compliance Integration: Includes a category-specific compliance rule library to flag high-risk words and suggest safer alternatives.
• Self-Audit System: Uses a five-dimensional scoring system (compliance, translation of benefits, visual impact, tactile expression, and narrative coherence) to trigger rewrites if scores fall below 80%.
• Platform Agnostic: Compatible with various Agent runtimes and AI tools like Claude Code, Cursor, and Windsurf.