kangarooking/director-skills
ResearchThis page was translated automatically. Technical details may differ from the original. View the original
Director Skills: Open-source Agent Skills for AI Video Creation | Director Skills: Open-source Agent Skills for AI video creation.
Overview
Director Skills: Let Agents organize creativity, scripting, storyboarding, generation, and final content into an executable AI video workflow like a director. It is an open-source Agent Skills set for AI video creation. Each skill is a dedicated "AI director," using structured workflows to help agents complete creative tasks from requirement understanding to deliverable results. Suitable for Claude Code, Codex, and other agents that support the Agent Skills open standard. Current status: Director Skill has included its first specialized Skill travel-skill. - Understand creative goals, audience, platform, duration, and production constraints - Complete creative positioning, narrative structure, visual style, and camera design - Output directly generated segment prompts, frame transitions, and consistency anchors - Check facts, copyright, cultural expression, model capability, and delivery boundaries - Evaluate visuals, continuity, and final film completeness through repeatable acceptance criteria When a Skill is marked as "Available," tools like Claude Code or Codex can directly say: 1. Includes valid frontmatter and complete SKILL.md. 2. Clearly state trigger scenarios, inputs, outputs, steps, and boundaries. 3. Provide installation and error handling instructions when tools or scripts are needed. 4. Run at least one real case and retain verifiable acceptance results. 5. Do not falsify generated results, nor hide model, platform, copyright, or data restrictions. New directoring skills, case studies, evaluation methods, and practical tools are welcome. Please read the CONTRIBUTING.md before you begin.
Overview (machine translated)
🎬 Director Skill · Director Skills
Director Skills: Let Agents organize creativity, scripting, storyboarding, generation, and final content into an executable AI video workflow like a director.
Director Skills (Director Skills) is an open-source Agent Skills set for AI video creation. Each Skill is a dedicated “AI Director,” using structured workflows to help Agents complete creative tasks from requirement understanding to deliverables.
Suitable for Claude Code, Codex, and other agents that support the Agent Skills open standard.
Current status: The Director Skill includes 7 specialized skills, all of which are available. The newly added
seedance-motion-studio(V0.1) has been confirmed by the author to complete testing, covering the planning and production workflow for brand promotional videos, product introductions, news commentary, and popular science videos.
📋 Skills
| Name | In a sentence | Status |
|---|---|---|
| 🎥 seedance-motion-studio | Organize requirements such as brand promotional videos into research, creativity, storyboarding, reference mapping, generation, editing, and acceptance processes | Available · V0.1 |
| 🏝️ travel-skill | Planning, writing, reviewing, and restoring cultural tourism promotional videos that blend real footage with AI footage | Available |
| 📷 travel-photo-to-video | Organize real travel photos into cultural and tourism storyboards, output a unified visual master and a per-shot first-frame image to video prompt | Available |
| 🎨 cinematic-asset-prompts | Extract characters, scenes, props, and vehicles from the script, and output bilingual Chinese-English visual asset prompts | Available |
| ⚔️ action-fight-prompt | Design cinematic action action cues with timing, action causality, environmental feedback, and asset locking | Available |
| 🎭 emotion-performance-director | Convert emotions, lines, and subtext into observable micro-performance cues that unfold over time | Available |
| 🎞️ miora-video-studio | Generate supplementary parameters for WorkBuddy/Miora videos Gate, task polling, order verification, and failure recovery | Available |
🔎 Skills Description
travel-photo-to-video
For users who already have 1–10 real travel photos and travel locations, analyze the subject and movable elements in the photos, recommend the sequence of shots, unify the visual style, and output independently copyable first frame to video prompts for each photo. Each frame is set to 4 seconds by default, with adjustments made within 3–6 seconds depending on the content, preserving the subject, architecture, space, and lighting relationships in the first frame. It is recommended to generate each shot before editing.
cinematic-asset-prompts
This page was translated automatically. Technical details may differ from the original.
Recommended Tools
Try a different keyword or remove a filter.
Install
npx skillfish add kangarooking/director-skills