Sobre o skill

Recreate a short vertical talking-head viral video (captions, tier list, B-roll pops, avatar) from a word-anchored timeline JSON, rendered deterministically with HyperFrames and FFmpeg, and measure the clone against the source frame by frame. Can break a source video down into a timeline draft, art-direct the new host and cards in one style, swap the A-roll and voice, keep a paid-call ledger, and build a side-by-side acceptance grid against a reference. Use when the user asks to clone, remake, rebuild or compare a short video's layout and rhythm with a new host, voice or images. Local CLI only for now; no hosted (Lab online) mode yet.

O que preparar

Forneça briefing e recursos. Revise, refine e exporte. Alguns skills exigem ferramentas ou modelos pagos.

Documento do skill