zhouwei713 /seedance-prompt

About seedance-prompt

Give it a vague scene. It picks who filmed it, on what device, and in which era. Then it writes device flaws, small mistakes, ambient sound and a 10-15 second timeline. The skill works for AI video models, and all docs and skill text are in Chinese.

Best for: Creators using Seedance, Sora, Kling, Runway or Veo who want handheld, documentary or archival-looking footage instead of clean AI samples.

• Splits your idea into 7 parts: character, place, style, camera, timeline, audio, goal
• Adds device looks like 2000s DV, VHS, Super 8, phone vertical and CCTV
• Turns vague moods into things you can film, like steam or flickering lights
• Writes a 10-15 second timeline with small, natural mistakes
• Runs a self-check for face consistency, audio sync and ad-style framing
• Ships SKILL.md plus camera-aesthetics.md, atmosphere-dictionary.md, anti-ai-checklist.md

1. 1
   Clone the repo into your Hermes skills folder: the command below

   git clone https://github.com/zhouwei713/seedance-prompt.git ~/.hermes/skills/creative/realistic-video-prompting

   Copy
2. 2
   Restart Hermes or start a new session, then invoke realistic-video-prompting
3. 3
   Without Hermes, drop SKILL.md into Claude, Codex, Cursor, OpenCode or another agent that supports skills or long prompts
4. 4
   Or read it as a method: SKILL.md, references/camera-aesthetics.md, references/atmosphere-dictionary.md, references/anti-ai-checklist.md
5. 5
   Describe a scene in plain words, e.g. a New York street in the morning with a real phone-snapshot feel, 15 seconds