What problem does it solve? Turning a full-length movie or long-form video into a narrated commentary cut requires watching the entire source, writing grounded narration, synthesizing voiceover, aligning subtitles, and rendering a final video. This Skill automates that pipeline as a durable, resumable project with state-based routing and strict QA validation. ## Core Features & Use Cases - Complete-film evidence building: Analyzes the source in chronological overlapping clips via Omni calls, producing auditable watch notes and measured execution facts (credits, subtitle band, canvas). - Grounded narration planning: Authors a validated editing plan and narration script with evidence-referenced movie locators, then freezes immutable shards for execution. - Resumable rendering and QA: Renders shards with TTS voiceover, ducked source audio, and measured subtitles, then assembles and validates the final video with decode, loudness, black/freeze, seam, and subtitle checks. - Use Case: Point the Skill at a local movie file and request a full Chinese-language commentary cut; it inspects project state, builds evidence, plans narration, renders shards, and delivers a QA-passed commentary.mp4. ## Quick Start Create a full narrated commentary video from my local movie file using this Skill, starting by inspecting the project state with the provided script.
