Cross-Border Product Selection Assistant

Problem

When evaluating whether a product is worth selling, cross-border e-commerce operators usually need to weigh demand signals, competition density, price bands, target-market differences, profit room, and risk factors. Relying only on marketplace rankings or gut feeling can miss cost assumptions and structured reasoning. This skill turns the process into a repeatable workflow: from a natural-language request to a full product-selection report, profit estimation, and library management.

Core Capabilities and Key Steps

• Product analysis: After a request such as “analyze whether this product is viable,” the workflow runs analyze, then local-model report generation, then deliver; the server handles authentication, task messages, report parsing, and card generation, while local inference performs the reasoning.
• Full reports: The expected output is not a short summary. It should preserve the opening, five-stage analysis, key data notes, and final recommendation, rather than only a quick-glance table or one-paragraph conclusion.
• Profit reverse calculation: The profit capability performs calculation-only estimation for target revenue and required sales volume, without relying on generative output for conclusions.
• Selection library: save and library let operators preserve evaluated products for later comparison and re-evaluation.

Scope and Limitations

This is useful for structured evaluation of target markets, profit potential, and product differentiation in e-commerce operations. It depends on an external selection API and local model inference; report completeness affects automatic library ingestion. If the report is below the required standard, a full draft is still output, but manual saving may be needed. Internal prompts, commands, or methodology details should not be exposed in the user conversation.