openmotion-shortlist-shorts
OpenMotion — Cut one long podcast, interview, course or livestream into stand-alone vertical shorts by building a scored shortlist first, then finishing one short for approval before the batch. Use when the user wants clips, highlights or shorts from a long recording.
Try your first task
A source-grounded starter task has not been verified for this skill yet.
SKILL.md
name: openmotion-shortlist-shortsdescription: >- OpenMotion — Cut one long podcast, interview, course or livestream into stand-alone vertical shorts by building a scored shortlist first, then finishing one short for approval before the batch. Use when the user wants clips, highlights or shorts from a long recording.OpenMotion · Shortlist Shorts
Build the video as a native OpenMotion HTML/CSS/JS composition through the installed OpenMotion production Skill: create or open the project, author the composition, save with revision protection and preview. This library Skill adds creative direction only; it registers no MCP Tool or production route. Treat palette, font and timing values below as starting points; a supplied Brandkit or the user's brand wins. For missing images, voice or music, first discover providers and model IDs with openmotion_media_use, then generate with openmotion_assets_use; get the user's consent before any paid generation. Never invent quotes, numbers, specs or dates.
Status: recipe / draft. No executable template or rendered comparison yet.
Use for: cutting one long podcast, interview, course or livestream into several stand-alone vertical shorts.
Method: shortlist before cutting
- Transcribe with word timing (host or provider capability). Do not guess words from frames.
- Build a shortlist table before editing anything: start–end time, the hook line quoted from the transcript, why it stands alone, a 0–100 score and any risk (needs earlier context, names a private person, unclear audio).
- Let the user pick. Default to offering at least four candidates when the recording supports them; if there are fewer strong moments, say so instead of padding.
- Finish one short first for style approval, then batch the rest with the same style.
Selection rules
| Rule | Starting value |
|---|---|
| Length | 20–75 s, never above 90 s unless asked |
| Count | Roughly one candidate per 5–10 minutes of source; total shorts no more than half the source length |
| Strong signals | A hook in the first 1–3 s, a clear payoff, emotion or contrast, a specific example or number |
| Red flags | Needs earlier context, table-of-contents summaries, cut-off endings |
Edit and format
Start on the hook sentence; remove filler and long pauses but keep natural breaths. Reframe to 9:16 after checking where the face or screen content actually is: face-follow crop for a single speaker, stacked layout for two speakers, full frame on a blurred background for screen recordings. Add word-timed captions in the lower third, clear of platform UI.
Review
Each short makes sense without the original; the first line is the hook; captions match the speech; crops never cut off the active speaker.
Provenance
Written for OpenMotion after surveying publicly listed workflow names and descriptions in ChatCut's preset gallery (November 2026). Original text; no ChatCut prompts, source code, templates or media are included. Apache-2.0, like the rest of the repository.