Short-Form Video and Video Production
Principles
- Each video should communicate only one core takeaway.
- Use the first 3 seconds to present a conflict, result, benefit, or concrete scenario.
- The voiceover must sound natural when spoken; do not simply turn a long article into a sequence of subtitles.
- Visuals should provide evidence, demonstrate, compare, or convey emotion rather than repeat the subtitles.
- Prioritize real human voices, product screen recordings, and real evidence; do not default to digital humans or animation.
- Digital humans are a production mode that users must actively choose. Rights to the likeness and voice are required. The user is responsible for manually uploading, creating, and selecting sensitive source assets on the chosen digital-human platform, while retaining manual preview and publishing confirmation.
Confirmed Inputs
Check: target platform, audience, content objective, duration, on-camera approach, available assets, core takeaway, call to action, and content that must not be shown. If missing information would change the plan, ask no more than 3 questions.
Video Production Modes
- : If there is no source footage, the readiness of the digital-human platform assets has not been confirmed, or there is no explicit authorization to produce the final cut, deliver a shoot-ready plan and do not generate a final cut without authorization.
- : When the user provides source footage and explicitly requests editing, use the video, subtitle, and audio capabilities available in the current environment.
Digital Human Production Mode
Animated-Explainer Production Mode
The implementation of animated-explainer mode is also discovered at runtime. It may be an installed animation-production Skill, an editing-software project template, or a user-specified video tool. The core workflow is not hard-coded to a particular image or voice service. Before any paid generation, explain the estimated cost to the user.
The digital-human generation tool is also discovered at runtime. If the current environment provides Topview MCP, it can act as an adapter for generating digital-human segments after the user confirms that the digital human and voice are ready in Topview. This Skill does not use upload credentials or cloning APIs to upload the user's original headshot or recording on the user's behalf. The core workflow is not hard-coded to a specific provider API. If generation or synthesis capabilities are unavailable, fall back to a digital-human storyboard, visual mapping, subtitle script, asset checklist, and a publishing package that can be completed manually.
Multi-Platform Versions
When a final cut is intended for multiple platforms, first choose a master version—typically 16:9 for YouTube and 9:16 for vertical platforms—then derive the other versions according to video-platform-versions.md: landscape-to-vertical adaptation (native re-layout or landscape-to-vertical packaging), burned-in subtitles or SRT files, platform-specific covers, AI disclosure labels, and music attribution. Use video-release-matrix-template.md to record each version and its publishing fields.
Structure
Choose one primary narrative structure:
- Problem → Method → Example.
- Result → Process → Lesson.
- Conflict → Choice → Consequence.
- Demonstration → Explanation → Scope of applicability.
End with one memorable takeaway and one primary call to action.
Storyboard
Use storyboard-template.md. For every shot, specify:
- Time and duration.
- Shot size, visual, subject action, or screen-recording operation.
- Voiceover, subtitles, and on-screen text.
- Evidence, B-roll, sound effects, BGM, and transitions.
The sum of all shot durations must equal the target duration. Facts in the subtitles and voiceover must be consistent.
Output Package
Deliver:
- Three hooks and the recommended version.
- A complete voiceover script.
- A storyboard or screen-recording checklist.
- A subtitle script.
- A cover title and visual direction.
- Publishing title, body copy, hashtags, and BGM suggestions.
- Filming or asset gaps.
Digital-human production mode additionally delivers: likeness and voice rights confirmation; status of manual platform uploads and asset readiness; the generated digital-human segment; visual analysis and subject-window specifications; a subtitle file or subtitle text; a pronunciation glossary; final video technical specifications; AI/digital-human disclosure requirements; and acceptance results. The delivery package does not include original headshots, audio recordings, or platform-private identifiers.
Animated-explainer production mode additionally delivers: the master version and platform-specific derivatives; subtitle files; covers for each platform; music licensing and attribution text; a pronunciation glossary; a fact-source list; technical specifications such as loudness and aspect ratio; the production project directory; and rebuild commands.
For multi-platform publishing, deliver a release matrix. Hand uploading and publishing off to
self-media-video-publisher
and comply with the authorization requirements for each platform.
If the user has not provided source footage, stop at the shoot-ready plan by default. However, if the user explicitly chooses digital-human production, completes the rights confirmation, and confirms that the digital human and voice have been manually uploaded and selected on the chosen platform, proceed to discover and invoke the digital-human, video, subtitle, and audio tools available in the current environment.