Video Creator

Documents

video-creator

Video Creator is an Agent Skill for documents published in the majiayu000/claude-skill-registry repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.

Orchestrates end-to-end video generation through sequential workflow steps (audio, direction, assets, design, coding). Activates when user requests video creation from a script, wants to resume video generation, mentions "create video", "generate video", or "video workflow", requests running a specific step (audio, direction, assets, design, coding), asks to "create audio", "generate direction", "create assets", "generate design", or "code video components", or wants to resume a video. Manages workflow state tracking and parallel scene generation.

View on GitHub ↗

QUICK START

How to use this skill

Bring this guide into your coding agent with a prompt tailored to the tool you use.

Open your project in Codex.
Copy and paste the prompt below.
Review files and risks before approving any action.

Prompt to paste

Inspect the Agent Skill at https://github.com/majiayu000/claude-skill-registry/blob/HEAD/skills/data/video-creator/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/video-creator, but wait for my approval before writing files or running scripts.

Copying this prompt does not install or run the skill. Review third-party files before use.

Video Creator Skill

Manages the complete video creation pipeline from script input to final video components.

Execution Modes

Determine which workflow to follow based on user input:

Mode 1: Execute Video Step

When: Immediately read when user wants to execute a specific workflow step (audio, direction, assets, design, or coding). Workflow: execute-video-step.md

Mode 2: Execute Regenerate Scenes Step

When: Immediately read when user wants to regenerate specific scenes for a step (design or coding) with scenes parameter. Workflow: execute-regen-scenes-step.md

Mode 3: Resume Video

When: Immediatly read when user wants to continue an existing video workflow from where it left off or check the status of an in-progress video Workflow: resume-video.md

Mode 4: Create Video

When: User wants to create a completely new video from scratch, starting with script input and style selection Workflow: create-video.md

Step References

Individual step implementations (referenced by execution modes):

user-input-step.md - Gather script and video style
audio-step.md - Generate audio from script
direction-step.md - Create video direction
assets-step.md - Generate SVG assets
design-step.md - Generate design specifications
code-step.md - Create video components
regenerate-design-step.md - Regenerate design for specific scenes
regenerate-code-step.md - Regenerate code for specific scenes

Utility Scripts

video-status.py: Manages workflow state and step progression

Available actions:

create-video-status-file: Initialize new video workflow and return first step
get-incomplete-step: Get current pending step
complete-step: Mark step done and advance to next

Output Format:

All actions output a single line

list-topics.py: Get all topics with video workflows

Usage:

python .claude/skills/video-creator/scripts/list-topics.py

When user doesn't specify a topic, use this script to get all available topics, display them to the user, and ask them to specify which topic they want to work with