video-workflowslisted

# Chitra · Video Workflows Select one recipe, then run `create-video` from intake through release. This skill chooses format and beat structure; it never bypasses Direction, gates, rendered evidence, or critique. ## Product launch Use `brand-film` or `product-demo`, normally 30–45 seconds. Structure: cold open from approved positioning → problem → turn → product proof with real UI → close. Acquire real UI before scoring. Copy comes from approved product truth, not invented claims. Use a music bed and sparse SFX under MO-AUD-3. ## Screen demo Use `product-demo`, normally 45–90 seconds. Start from a screen recording and optional narration. When speech or clip audio matters, follow create-video's `footage-editing.md`: lock the provider-neutral word transcript, read the compact phrase pack, approve an explainable EDL, and render a source-audio-preserving plate. Before rendering, request evidence only for the selected EDL segments; inspect word-labeled filmstrips, waveforms, and adjacent cut strips for gestures, UI continuity, attention handoff, rhythm, and audio discontinuity. Select 3–6 timestamped proof moments, use moving clips for interaction and stills for holds, and crop recorder chrome, personal data, toolbars, and transient toasts before scoring. Captions describe what the footage actually proves. ## Social short Use `social-short`, normally 9:16 at 1080×1920 for 8–20 seconds. Keep one idea: hook within two seconds → one proof beat → mark. Use the register type min