What problem does it solve? Turning raw footage and photos into a polished vertical reel normally requires manual editing, caption writing, voiceover syncing, and format conversion. This Skill automates that pipeline: it probes and validates your assets, arranges them into a beat sheet, cuts each clip to its beat with FFmpeg, overlays kinetic text in an HTML composition, and renders a deterministic 4K MP4 with a synced voiceover and platform-ready caption pack. ## Core Features & Use Cases - Asset gate and auto-arrangement: check-assets.mjs probes every video/image with ffprobe, validates it against the 9:16 spec, aliases files, and drafts a storyboard.json so the best asset lands on the hook beat. - Deterministic cutting and rendering: cut-assets.mjs produces exact-length muted 1080x1920 clips per beat, and render-frames.mjs seeks a paused GSAP timeline frame-by-frame via Playwright to output a 2160x3840 MP4 with muxed audio. - Caption pack and audit harness: generate-caption.mjs writes multi-platform captions in the 500-900 character window with zero hashtags, and audit-asset-reel.mjs plus a fresh auditor subagent score the reel (/50) before delivery. - Use Case: Drop product photos and clips into an assets folder and ask for a 15-second aesthetic reel — the agent validates assets, cuts beats, adds a Kokoro voiceover, renders 4K, and ships captions. ## Quick Start Ask your AI agent to use the video-asset-reels skill to make a 15-second reel from the clips in your assets folder with a hook, per-beat captions, a voiceover, and 4K output.
