Seedance 2.0 — Shot Design

What it does

This skill provides a comprehensive prompt-engineering framework for Seedance 2.0, acting as a virtual film director that converts user ideas into structured, high-quality video prompts. It includes language rules, a cinematography dictionary, director style presets, a three-layer lighting/quality anchor system, and a validation checklist to ensure prompts meet platform limits and safety rules. Use cases include text-to-video, image-to-video (I2V) workflows, multi-modal reference handling, character swaps, dialogue-driven short films, and music-synced edits.

When to use it

Use this skill when you need production-ready prompts for Seedance 2.0 — for example, producing short cinematic scenes, converting a storyboard into a time-stamped prompt, extending or restyling existing video clips, or generating multi-segment prompts for videos >15s. It is intended for both Chinese and non-Chinese users and automatically selects the output language and formatting.

What's included

• Scripts: none (instruction-only skill).
• References: includes in-repo guidance files (cinematography, quality anchors, examples) when present.
• Instructions: step-by-step workflow covering requirement parsing, pre-production diagnostics, six-element prompt assembly, strict validation rules (length, camera language, asset limits, conflict checks), and final delivery templates in both Chinese and English. The skill also documents multi-segment assembly and IP compliance strategies.

Compatible agents

Inferrable: agents or toolchains that support prompt engineering for text-to-video workflows (Claude-style instruction agents, multimodal assistant workflows, or custom pipeline runners for Seedance).