AgentMatter
EN/中文
SubmitGitHub
Skills · Community project

seedance2-skill

A bilingual prompt-writing Skill for Jimeng Seedance 2.0 covering multimodal reference syntax, camera language, timed segments, sound direction, and templates for ads, dramas, music videos, and education.

3.4k StarsMITMarkdownUpdated 6 months ago
Install and use GitHub
01

Project overview

A bilingual prompt-writing Skill for Jimeng Seedance 2.0 covering multimodal reference syntax, camera language, timed segments, sound direction, and templates for ads, dramas, music videos, and education. Its main documented capabilities are Reference-asset orchestration, Professional camera language, Timed prompt segments, Scenario templates. This page is based on public GitHub material and does not present repository claims as hands-on testing.

02

Core capabilities

01

Reference-asset orchestration

Assigns `@`-referenced images, video, and audio to first-frame, character, scene, camera, rhythm, or sound roles.

02

Professional camera language

Structures subject, scene, action, camera, transitions, sound, and style into a clear prompt.

03

Timed prompt segments

Breaks longer videos into timed opening, development, climax, and resolution segments.

04

Scenario templates

Provides prompt patterns for ads, short dramas, music videos, ecommerce, and education.

03

Installation and usage

Install the repository through the Skills CLI or manually copy the English `SKILL.md` and Chinese `zh/SKILL.md`. Then ask the agent to turn a simple idea into a structured prompt.

AI AGENT INSTALL

Let an AI Agent install it

Send this prompt to Codex, Claude Code, or another AI agent that can work with your local environment.

Help me install seedance2-skill from https://github.com/dexhunter/seedance2-skill. Read README.md, README-zh.md, SKILL.md, zh/SKILL.md, and LICENSE first, run `npx skills add dexhunter/seedance2-skill`, and confirm how the English and Chinese Skills are discovered. Ask before replacing a Skill or file. Then create a prompt for a ten-second product showcase without sensitive real-person assets, checking subject, scene, action, camera, timed segments, sound, style, and explicit roles for every `@` reference. Do not log in to Jimeng, upload assets, or generate video without permission.
01Before you start
  • A host that supports Agent Skills
  • Select `zh/SKILL.md` for Chinese guidance
  • Actual video generation requires separate access to Jimeng Seedance 2.0
02Copy the install command or configuration
npx skills add dexhunter/seedance2-skill
03Complete the setup steps
  1. 1
    Review prerequisites and boundaries

    Read the repository installation material and confirm: A host that supports Agent Skills; Select `zh/SKILL.md` for Chinese guidance; Actual video generation requires separate access to Jimeng Seedance 2.0. Stop for confirmation before credentials, cost, additional permissions, or replacement.

  2. 2
    Run the documented installation

    Use the current README command: npx skills add dexhunter/seedance2-skill

  3. 3
    Run a minimal verification

    Ask the agent to write a ten-second product showcase prompt containing subject, scene, action, camera, timed segments, sound, and style, and verify that every `@` reference has an explicit role.

  4. 4
    Move to a real task

    After the minimal check passes, use nonsensitive test media first. Ask separately before external uploads, billable calls, publishing, or irreversible changes.

How to verify the setup

Ask the agent to write a ten-second product showcase prompt containing subject, scene, action, camera, timed segments, sound, and style, and verify that every `@` reference has an explicit role.

Before using it
  • The project writes prompts; it does not include the Seedance model, a Jimeng account, or a video-generation API
  • Model parameters and platform limits can change, so recheck current official UI and documentation before submission
  • Do not reference people, voices, brands, or copyrighted assets without authorization
  • No Jimeng login or video generation was performed
04

Use cases

SCENARIO 01

Product and ecommerce prompts

Combine product images, scene references, and ad pacing into a shot-oriented prompt.

SCENARIO 02

Short dramas and music videos

Design character action, camera, transitions, dialogue, or beats across timed segments.

SCENARIO 03

Reference-guided recreation

State precisely whether each image, video, or audio reference contributes camera, action, style, rhythm, or sound.

05

Assessment

Based on the README, installation material, and repository structure, seedance2-skill turns Reference-asset orchestration, Professional camera language, Timed prompt segments into explicit agent workflows. Its limitations are mainly dependency, host, and data-boundary concerns. This assessment is based on public material, without installation or a real media task.

Why it may be useful

  • Assigns `@`-referenced images, video, and audio to first-frame, character, scene, camera, rhythm, or sound roles.
  • Structures subject, scene, action, camera, transitions, sound, and style into a clear prompt.
  • Breaks longer videos into timed opening, development, climax, and resolution segments.

What to know first

  • The project writes prompts; it does not include the Seedance model, a Jimeng account, or a video-generation API
  • Model parameters and platform limits can change, so recheck current official UI and documentation before submission
  • Do not reference people, voices, brands, or copyrighted assets without authorization
  • No Jimeng login or video generation was performed
06

README

seedance2-skill

Overview

A bilingual prompt-writing Skill for Jimeng Seedance 2.0 covering multimodal reference syntax, camera language, timed segments, sound direction, and templates for ads, dramas, music videos, and education. skill to create best prompts for generating videos with seedance2.0

Getting started

  • Install the repository through the Skills CLI or manually copy the English `SKILL.md` and Chinese `zh/SKILL.md`. Then ask the agent to turn a simple idea into a structured prompt.
  • Project focus: A bilingual prompt-writing Skill for Jimeng Seedance 2.0 covering multimodal reference syntax, camera language, timed segments, sound direction, and templates for ads, dramas, music videos, and education.
  • Published characteristics: English and Chinese Skill files; Seedance 2.0 multimodal reference syntax; Camera and timed-segment templates; MIT license.
  • Ask the agent to write a ten-second product showcase prompt containing subject, scene, action, camera, timed segments, sound, and style, and verify that every `@` reference has an explicit role.

Configuration

npx skills add dexhunter/seedance2-skill
Read the complete README on GitHub →