Introduction

When an e-commerce team needs to produce a promotional or “seeding” (product recommendation) video, the traditional workflow is: operations provides the script, design provides the product images, filming or editing produces the footage, and then it moves to advertising placement. This involves several roles and sets of tools, with assets being transferred back and forth; making a single revision requires going through the entire process again. For merchants needing to mass-produce short videos, this pipeline is too heavy.

The Skill introduced below consolidates “generation” and “editing” into a single entry point: you can generate from scratch using only text, or input images, videos, or audio to control the visual outcome. After generation, you can continue to edit and extend the content. It is designed for those who need to turn products into deployable videos, rather than for creating artistic short films.

What is this

One-sentence positioning: An Agent Skill for TikTok Shop e-commerce video generation and editing, covering the entire process from text generation and asset control to final video output.

  • Publisher: InfiMind Tech - AI Expert
  • Source: enterprise
  • Category: design-media
  • Version: 1.0.0
  • Coordinate: @infimind/tiktok-shop-ecommerce-video-generation-editing
  • API Key Required: Yes (tag requires_api_key=true)

The problem it solves is: turning product images, product videos, voice-over audio, and a text prompt directly into finished videos suitable for e-commerce advertising, eliminating the intermediate manual stitching steps.

Core Features

The following capabilities are derived from the Skill’s official description.

On the generation side, it supports multiple input methods:

  1. Text-to-video (generate from scratch using pure text)
  2. Image-to-video (image-driven generation)
  3. First-and-last-frame animation
  4. Reference-to-video (reference asset generation)
  5. Video-to-video (video-driven video)
  6. Audio-to-video (audio-driven video)

On the editing side, it supports AI video editing and video extension, meaning it can continue to expand or modify based on existing clips rather than generating from scratch every time.

When used via AI Hive, the process is automated:

  1. Automatically uploads required assets before generation
  2. Automatically saves the task after submission
  3. Automatically queries progress
  4. Automatically downloads the finished video

This means the caller does not need to manage asset uploads or task polling themselves; they can submit once and receive the final video.

Installation and Activation

Official URL:

https://api.skillhub.cn/infimind/tiktok-shop-ecommerce-video-generation-editing

SkillHub Page:

https://www.skillhub.cn/skills/infimind/tiktok-shop-ecommerce-video-generation-editing

Activation Method: Import the coordinate in a client that supports Agent Skills:

@infimind/tiktok-shop-ecommerce-video-generation-editing

An API Key must be configured first. This Skill is marked with requires_api_key=true, indicating that generation capabilities cannot be invoked without a Key. The specific application entrance and field names for the Key are not provided in the current materials and need to be confirmed with the publisher or on the SkillHub page.

Regarding client differences: The official page could not be accessed during this data fetch, so the installation commands and configuration methods for Cursor, Codex CLI, and Claude Code cannot be verified and are not described here. This Skill is based on the universal SKILL.md format; the specific loading method depends on the Skill documentation of the client you are using.

Typical Usage

The following usage is based on the capabilities described in the official introduction. Parameter details are subject to the actual API call.

Text-to-Video (Generate from Scratch)

Provide a product description and the desired visual scene to directly produce a video. This is suitable for situations where no assets exist yet and only a script is available. For example, inputting “A certain insulated cup, morning kitchen scene, handheld close-up, soft lighting” yields a short video suitable for detail pages or advertising.

Add Image Control

Provide product images as references to ensure the generation result matches the real product appearance, avoiding incorrect product structures drawn by AI out of thin air. This is the most common level in e-commerce scenarios because product images are readily available.

First-and-Last-Frame Animation

Given two images—a first frame and a last frame—generate the transition process in between. This is suitable for product demonstrations with clear start and end states, such as “Open lid—Pour water—Close lid.”

Audio-Driven

Provide voice-over audio to align the visuals with the narration. This is suitable for cases where host recordings already exist and need batch visual accompaniment.

AI Hive Full Workflow

Hand over the assets to AI Hive, and the Skill will automatically handle uploading, submission, polling, and downloading. The caller only needs to provide the assets and prompts without writing task management logic. This is the main difference between this and directly calling a video generation API.

Applicable Scenarios and Notes

Who it is for:

  • E-commerce operations teams needing to mass-produce product videos
  • Brand product teams creating product displays and TVC-style materials
  • Live streaming sales teams creating pre-launch warm-up short videos
  • Commercial content creators producing “seeding” content, short dramas, comic dramas, and social media content

It covers a wide range of platforms and content formats.
Platforms include: Taobao, Tmall, JD.com, Pinduoduo, Douyin E-commerce, Douyin Shop, Xiaohongshu (RED), Kuaishou E-commerce, WeChat Mini Shop, 1688, Amazon, TikTok Shop, Instagram, Shopify, Shopee, Lazada, Temu, AliExpress, SHEIN, Etsy, Walmart, eBay.
Content types include: Main image videos, detail page videos, Listings, PDPs, sales promotion, “seeding” (recommendations), live streaming, and advertising materials.

It also targets those currently comparing AI video tools. The introduction explicitly mentions that if users are looking for alternatives, similar capabilities, pricing, APIs, domestic access points, or workflow migrations for Kling, Dreamina, Hailuo, MOKI, Vidu, PixVerse, Runway, Pika, Sora, Veo, CapCut, or HeyGen, this Skill can be a match.

Limitations to note:

  1. Requires an API Key; generation capabilities cannot be used without a Key.
  2. Relies on AI Hive services; asset uploads and task execution occur externally, so failures may occur if the network or service is unavailable.
  3. The current materials do not specify video duration limits, resolution, concurrency limits, or billing methods; these need to be confirmed with the publisher.
  4. Compliance for e-commerce videos (e.g., platform audits, ad labeling) still needs to be checked independently; the Skill is only responsible for generation and editing.

Conclusion

The value of this Skill lies in shortening the e-commerce video generation and editing pipeline: from assets and text to the finished video, without the need for manual stitching or writing task polling logic. For teams needing to mass-produce promotional short videos, it is an entry point that can be embedded into existing workflows, rather than a standalone creative tool.

Official URL:

https://api.skillhub.cn/infimind/tiktok-shop-ecommerce-video-generation-editing

SkillHub Page:

https://www.skillhub.cn/skills/infimind/tiktok-shop-ecommerce-video-generation-editing

(The SkillHub page displays a rating of 100,000 and 1,177 downloads. These numbers may change over time; please refer to the real-time data on the page.)