This connects Claude to LibTV's AI video platform through natural language workflows. You send conversational requests like "make an anime video" or "replace the boat with a heart" and the backend agent handles model selection, prompt engineering, and multi-step orchestration automatically. It's designed for complex tasks that would normally require manual workflow setup: generating complete short films from one sentence, cloning video styles for ads, creating music videos, or building storyboards. Your job is pure coordination: upload reference files if needed, pass the user's exact words to the API, poll for results, download when ready. The key insight here is staying out of the creative loop. Don't rewrite prompts or decompose tasks yourself. The backend knows the models better than you do.
npx skills add https://github.com/libtv-labs/libtv-skills --skill libtv-skill