What is this?
This is a 'Seedance 2.5 Video Director' Skill for Codex, built for the Dreamina (即梦) Seedance 2.5 video platform. It turns text ideas, character reference images, videos, audio, storyboards, or old prompts into ready-to-use video scripts, director plans, diagnostics, and copy-paste prompts.
This is a 'Seedance 2.5 Video Director' Skill for Codex, built for the Dreamina (即梦) Seedance 2.5 video platform. It turns text ideas, character reference images, videos, audio, storyboards, or old prompts into ready-to-use video scripts, director plans, diagnostics, and copy-paste prompts.Is this a video generator itself?No, it's a prompt and script generation Skill, not a video generator — it never calls paid generation APIs or consumes Dreamina credits. It only produces planning text, prompts and diagnostics that you paste into Dreamina yourself.
No, it's a prompt and script generation Skill, not a video generator — it never calls paid generation APIs or consumes Dreamina credits. It only produces planning text, prompts and diagnostics that you paste into Dreamina yourself.How do I install this Skill?If you have Node.js, run the one-line npx installer shown in the README from a Windows, macOS or Linux terminal: it fetches the Skill from GitHub and installs it into your Agent Skills directory with -g for a global install. A manual PowerShell (Windows) or shell (macOS/Linux) install is also provided.
If you have Node.js, run the one-line npx installer shown in the README from a Windows, macOS or Linux terminal: it fetches the Skill from GitHub and installs it into your Agent Skills directory with -g for a global install. A manual PowerShell (Windows) or shell (macOS/Linux) install is also provided.Do I need Node.js to install it?Node.js is needed for the recommended npx installer, but if you don't have it you can trigger installation directly by sending a message inside a Codex session instead.
Node.js is needed for the recommended npx installer, but if you don't have it you can trigger installation directly by sending a message inside a Codex session instead.Does it guarantee consistent video results?No — the underlying generation model is random, so the Skill's prompts can only reduce identity drift, physics errors and camera conflicts, not guarantee the exact same result every time. It also never submits generation jobs itself, so you still run the final prompt in Dreamina manually.
No — the underlying generation model is random, so the Skill's prompts can only reduce identity drift, physics errors and camera conflicts, not guarantee the exact same result every time. It also never submits generation jobs itself, so you still run the final prompt in Dreamina manually.What character reference rules does it follow?By default a reference image represents the whole visible person: identity, facial features, skin tone, hairstyle, height, body shape, visible clothing and accessories are kept, while background, pose, composition and lighting from the reference are not inherited unless you ask for a face-only reference.
By default a reference image represents the whole visible person: identity, facial features, skin tone, hairstyle, height, body shape, visible clothing and accessories are kept, while background, pose, composition and lighting from the reference are not inherited unless you ask for a face-only reference.What video lengths and editing modes does it support?It covers 4-30s normal clips with a precise 30-second timeline, 30-180s Long Video, pre/post continuation of existing clips, Smart Edit with marked-region or uploaded-video editing, seamless transitions between two clips, and multi-panel storyboard animation.
It covers 4-30s normal clips with a precise 30-second timeline, 30-180s Long Video, pre/post continuation of existing clips, Smart Edit with marked-region or uploaded-video editing, seamless transitions between two clips, and multi-panel storyboard animation.How is the Skill's quality checked?A local validator, also run by GitHub Actions on every push and pull request, checks the Skill's front matter and required files, install-guidance triggers, the agents/openai.yaml invocation name, and the behavioral test cases in tests/cases.json.
A local validator, also run by GitHub Actions on every push and pull request, checks the Skill's front matter and required files, install-guidance triggers, the agents/openai.yaml invocation name, and the behavioral test cases in tests/cases.json.Is this an official Dreamina or ByteDance product?No, the README explicitly states it is not an official ByteDance, Dreamina or 即梦 project; it's a third-party Skill that only produces planning text and prompts for use on those platforms.
No, the README explicitly states it is not an official ByteDance, Dreamina or 即梦 project; it's a third-party Skill that only produces planning text and prompts for use on those platforms.What are similar projects?h3-storyboard-skill, seedance2-skill, short-drama-production, film-studio-skills, manju-laoli-skill, awesome-seedance, lanshu-create-ai-presenter-video, hypit
h3-storyboard-skill, seedance2-skill, short-drama-production, film-studio-skills, manju-laoli-skill, awesome-seedance, lanshu-create-ai-presenter-video, hypitWhich topics is it in?seedance, storyboard, ai-video, video-editing, codex-skill, github-actions, video-generation, prompt-engineering, audio, testing, generative-ai, agent-skills
seedance, storyboard, ai-video, video-editing, codex-skill, github-actions, video-generation, prompt-engineering, audio, testing, generative-ai, agent-skills