← ClaudeAtlas

zobvialisted

Four modes plus bundled sub-skills. LEARN: watch a video and extract its editing style into a reusable style skill — use when the user says 'watch this', 'learn this style', 'analyse this video', pastes a video link (YouTube, Shorts, Reels, TikTok) or gives a local video file without an SRT. CREATE: turn an SRT/subtitle file into a Remotion motion-graphics video (1920x1080, 30fps, word-by-word animation, SFX only, full Bengali support) — use when the user says 'make a motion graphics video from this SRT', 'edit in <style> style', pastes or attaches an .srt file, or gives both a video and an SRT (learn the style, then create in it). DOCUMENTARY: turn a documentary title or topic into a complete 10-minute VOX-style YouTube documentary production package (research, script, 8-second scene map, text-to-video prompts, audio package, SEO packaging, thumbnails, checklist) in 10 approval-gated stages — use when the user says 'documentary', 'ডকুমেন্টারি', 'make a documentary about', or gives a documentary title/topic.
zobaerio/zobvia · ★ 1 · Data & Documents · score 74
Install: claude install-skill zobaerio/zobvia
# Zobvia One skill, four jobs: **learn how a video is edited**, **create a new video in a learned or built-in style**, **plan a full documentary production package** from a title or topic, and **edit real footage** with FFmpeg. Style is learned; footage, voice, and music are never reused. Bundled sub-skills extend it further: a 3D cartoon explainer engine (Part E), the ffmpeg-skill toolset (Part F), a headless-Blender 3D asset toolkit (Part G), a typed audio-production engine (Part H), a colour-grading engine (Part I), an HTML-scene video renderer (Part J), a production BRAIN that plans and orchestrates the others (Part K), a Remotion-on-footage bridge (Part L), a thumbnail renderer (Part M), a media-analysis engine (Part N), a QC checker (Part O), a subtitle engine (Part P), a transcription engine (Part Q), a typed video-editing graph executor (Part R) and a deterministic motion-graphics renderer (Part S), face/text/object detection (Part T), AI dubbing/narration (Part U), speaker diarization (Part V), verified media ingest (Part W), an artifact verification engine (Part X), an ImageMagick still-image toolkit (Part Y) and a native Auto-Shorts engine (Part Z). ## Philosophy - **Measurable numbers over adjectives.** Not "fast-paced" — "median shot 1.8 s, 34 cuts/min, captions 3 words/line". Every claim in a style skill or a plan must be a number, a frame count, or a hex value. - **Never guess.** If a tool, file, or measurement is missing, say so and stop that step. Do