What problem does it solve? Long-form podcast and video episodes contain shareable moments, but manually finding them, cutting clips, and writing platform captions is slow and inconsistent. This Skill automates the pipeline from transcript to ready-to-post vertical shorts. ## Core Features & Use Cases - Virality scoring: clip-finder.mjs parses a timestamped transcript and ranks moments by hook strength, emotion, controversy, quotability, and standalone value, outputting a scored clip-plan.md. - Automated cutting: Generates FFmpeg commands for 9:16 vertical crops at 1080x1920 and can execute them directly against the source video. - Hooks and captions: Produces a 2-second on-screen hook plus per-platform captions (TikTok, Reels, Shorts) with hook-first structure, one CTA, and zero hashtags. - Audit gate: audit-clips.mjs runs automated checks and a fresh auditor subagent completes a 50-point clip-worthiness scorecard before delivery. - Use Case: Feed a one-hour podcast transcript and episode video, then receive five scored clips, caption packs, and a signed audit report ready for posting. ## Quick Start Ask the agent to turn your podcast episode transcript and video file into five viral vertical shorts with hooks, captions, and FFmpeg-cut clips.
