What problem does it solve? Producing short-form video with AI normally means juggling separate tools for copywriting, storyboarding, image generation, voiceover, subtitles, animation, and final assembly, with character and scene consistency easily lost between sessions. This Skill turns a topic into a finished short video through one deterministic pipeline where assets stay reusable, traceable, and versioned across an entire series. ## Core Features & Use Cases - End-to-end production state machine: Guides copywriting, storyboard linting, character/prop/scene reference sheets, image generation, voice casting and TTS, animatic review, video generation, and final assembly, with mandatory human confirmation at each starred node. - Three render modes and multi-aspect output: Supports kenburns (zero-cost stills), dubbed (fixed-voice lip-synced image-to-video), and native (model-native audio-video) modes across 16:9, 9:16, and 1:1 canvases. - Consistency and cost guardrails: Enforces reference-sheet binding for characters and scenes, dry-run cost review before any paid video generation, review/version-stack management, and pre-delivery lint, verify, and consistency checks. - Use Case: Give it a topic like a book review or a short drama episode; it produces the script, shot list with bilingual prompts, character sheets, storyboard images, voiceover, subtitles, and a finished assembled video under a single project directory. ## Quick Start Ask the agent to turn your topic into a short video, for example: "Use kinema to make a 60-second narrated video about the Pomodoro Technique, starting with the script and storyboard for my review."
