AI E-commerce Listing Image Set

Specific Problem Addressed

When listing products on e-commerce platforms, sellers need to prepare a set of images for each SKU, including main image, selling points, details, scenes, etc., to attract buyers and meet platform standards. However, common challenges arise: product photos inconsistent with SKU facts, fabricated selling points not verified from user materials, main image compliance issues, lack of unity across images, or repetitive creativity. This leads to time-consuming, error-prone listing processes, impacting product exposure and conversion. This skill provides a systematic solution, focusing on generating an ordered, consistent listing image set for a single SKU.

Core Capabilities and Workflow

This skill creates images based on verified product photos and SKU information (e.g., name, model, color, material, size, packaging list) through a structured workflow. Key steps include:

• Input Confirmation and Planning: Collect hard inputs like product photos and confirmed SKU facts. No selling points are fabricated if facts are missing. First, output an ordered shot list (main image, selling points, scenes, sizes, packaging), detailing the purpose of each image, retained details, short copy boundaries, and canvas selection. Proceed to paid production only after seller confirmation.

• Per-Image Production and Tool Calls: Use the bundled scripts/mcp_client.py for uploads and remote calls. For new image generation, call beatra.images.transform with the confirmed product image in images[0]; for local corrections, use beatra.images.edit. Query beatra.models.list to decide on models, canvases, and pricing, with confirmation before decisions. Each image gets a unique client_request_id, default model as auto, and count as 1. Planning is free before production, and each paid task requires explicit approval with details on prompts, base images, etc.

• Review, Recovery, and Task Management: After production, check product shape, color, labels, fact consistency, image order, and canvas fit. Track tasks via beatra.tasks.get to completion; if responses are lost, use beatra.tasks.list to find candidates and confirm with beatra.tasks.get. Task cancellation is only upon seller request and does not replace already executed tasks. All calls must send JSON via standard input and cannot configure host Beatra Connector or downgrade to REST/OpenAPI.

Applicability and Constraints

• Scope: This skill is designed specifically for creating a unified multi-image set for a single SKU, ensuring each image consistently presents the same verified facts and brand direction. Main image compliance and supplementary image persuasion are handled separately, but integrated into an ordered set.
• Non-Applicable Scenarios: For new photography or broad creative exploration, use product-photo-studio; for checking single main image compliance for a specific site, use marketplace-main-image-preflight. Inputs must be real product photos and confirmed facts; no fabrication beyond user materials is allowed.
• Technical Constraints and Billing: Only the bundled MCP client is used; REST/OpenAPI downgrade is prohibited. Automatic updates check silently every 24 hours, and failures keep the current installation functional. Billing is per paid task, recording billing.net_charged_credits, with balance and error details in related documentation. Task recovery only replays the same request ID; any change in images, prompts, or canvas constitutes a new paid request.