Seedance 2.5 OS skill
A skill that turns a video idea into Seedance 2.5 settings, reference roles and one paste-ready prompt for the video model.
- 107 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- The skill writes only prompts and settings and creates no video
- The installer copies files into the personal skills directory
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/seedance-2-5-osDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skill writes only prompts and settings and creates no video
- The installer copies files into the personal skills directory
This is third-party code. Review the repository files before installing.
What it does
Seedance 2.5 OS rebuilds video work for version 2.5 instead of copying 2.0 recipes. The skill sorts a request into routes: a short clip, a 30-second story, an ultra-long narrative, image to video, multimodal references, extending an existing clip, a precise edit, green screen and storyboards. For each it picks a mode, duration, reference roles and prompt structure. For multimodal input the skill pins roles: `@Image1` controls identity, `@Video1` only motion, `@Audio1` rhythm. For an edit it limits the area and time interval, for example "from 0 to 8 seconds change only the blue trousers inside the red box". Separate parts describe repairing typical failures. A community project, not a ByteDance, Jimeng or Dreamina product.
Who it is for. Short-video authors, advertisers and editors who work with Seedance 2.5 and want precise prompts without manual tuning.
Good fit when
- A two-word idea must become a ready 10 or 30-second prompt.
- Images, video and audio must be tied to one clip with clear roles.
- A finished piece needs extending while keeping motion and camera.
- One detail in a finished clip is fixed without recreating the scene.
- A composite needs green screen shooting or a white model.
Not a fit when
- You need the video generator itself: the skill only prepares prompt and settings.
- Seedance 2.0 or another model is used: recipes target 2.5.
- Ready prompt collections are required: this is a method, not a catalogue.
Example request
Write a 30-second prompt: a vertical story about a girl adopting a stray cat on a rainy night.Limitations
The skill does not generate video, it prepares the prompt and settings for the Seedance service, access to which is needed separately. Figures about model capabilities such as 30 images, 10 videos and 10 audio files per request come from the project description and may differ from access in your region or plan. The `install.sh` installer copies the skill into the personal directory and backs up an older version. Documentation is in English with a Chinese translation in `docs/README.zh.md`. The repository was last updated on 2 August 2026 and uses the MIT licence.
How to disable. Delete the `seedance-25` folder from the personal agent skills directory. Remove timestamped backups the installer created by hand.
FAQ
Is this an official ByteDance product?
No. The README says it is a community project unaffiliated with ByteDance, Jimeng and Dreamina.
Which routes does the skill support?
Short clips, 30-second stories, ultra-long narratives, image to video, multimodal references, extension, edits, green screen and storyboards.
How are reference roles assigned?
Through tags: `@Image1` sets identity, `@Video1` only motion, `@Audio1` rhythm.
How to install on Windows?
Run `install.ps1` from PowerShell; on macOS and Linux use `./install.sh`.
What if generation fails?
The skill has sections on repairing typical failures: a local fix first, then regenerating only the problem piece.
Discussion
No questions or reviews yet.
Related
SkillsA skill that teaches an agent to write prompts for ByteDance's Jimeng Seedance 2.0 video model: @ references, camera language, timing, templates
A dialogue that turns an idea into a storyboard and a prompt for the Seedance 2.0 video model in Jimeng, with six templates and ten examples
A skill that improves Seedance prompts, plans storyboards and generates video through Atlas Cloud, plus a library of ready prompts
A Codex skill that plans fight rhythm, camera and effects in Seedance 2.5 prompts and revises the prompts you already have.
A skill that drives AI video through a clay-render scene: intake, geometry checks before render and a prompt from the same data.
A skill that takes a product name to a 30-second Apple-keynote-style video for Seedance 2.5, with keyframes and a quality check.
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| Seedance 2.5 OS (this card) | no | 107 | |
| lengyi-seedance2.5-prompt | no | 147 | turns an idea into a multi-shot 30-second prompt; here task routes, extension, edits and reference roles |
| Awesome Seedance 2 Prompts | no | 2,217 | a collection of 2000 prompts for version 2.0; here a method of building prompts for 2.5 |
README in short
The README compares 2.5 capabilities with 2.0, gives a table of requests by goal and describes installation scripts for macOS, Linux and Windows.