What problem does it solve?
When an AI video generation step (such as seedance) is blocked by content moderation, a downstream video merge step is left without a clip for that shot. This Skill turns the already-generated still image into a valid MP4 with a slow Ken-Burns zoom and a silent audio track, so the merge pipeline can still produce a complete deliverable.
Core Features & Use Cases
- Still-to-Video Conversion: Wraps ffmpeg's zoompan filter to produce an H.264 MP4 at a configurable resolution, duration, frame rate, and zoom rate from a single PNG or JPG.
- Silent Audio Track: Adds an AAC silent audio stream so merge steps that demux or remux audio do not fail on a missing stream.
- Moderation Fallback: Designed as an on_failure substitute in meta-skill DAGs, replacing a blocked AI video generation step with a placeholder clip that preserves character, framing, and timing.
- Use Case: In a short-drama production pipeline, when seedance refuses to animate shot 3 due to a flagged face, run this Skill on the existing shot 3 PNG to emit a 5-second 720x1280 clip that the merger can consume without manual intervention.
Quick Start
Ask the agent to turn the still image shot1.png into a 5-second 720x1280 MP4 named shot1.mp4 using this Skill as the fallback for the blocked video generation step.
