Capture raw content (metadata, transcript, frames) from YouTube videos and Instagram reels, analyze it in the current Claude session, download YouTube videos (full or trimmed), and extract multiple clips from long-form content (podcasts, courses) via a conversational brainstorm. Skills: view-youtube-video, view-instagram-reel, extract-content, clip-youtube-video, clip-video-parts.
Screen demos and B-roll for creators, from a script or plain instructions: screen (real headed Chrome, one or two windows, beats timed to a voice or to seconds, the pointer moving like a hand), assemble (large Screen Studio-style pointer, spring zooms, plain browser frame, fades, 4K 60 fps, normalized loudness), audio (loudness, peaks, pauses and speech rate; normalize to -16 or -14 LUFS; pause cap), voice (optional: generate with ElevenLabs, or bring your own audio or video and get line and word timings; shorter pauses), broll (optional, paid: people B-roll with Higgsfield). macOS, Linux and Windows.
WhatsApp Status content generator using Nano Banana AI - image templates, video generation, scheduling reminders. Create professional Status content for Kenya businesses.
Writing for videos and content. Skills: hooks for the opening (six formats, the spoken, text and visual hook lined up, the lock-in zone, four mistakes), storytelling for the structure that keeps people watching (stakes, a big question, head fakes, re-hooks), and video-script for scripts recorded one take per line.
Produce deep, long-form blog posts (1,800–2,500+ words) by interviewing the user first, researching the topic, architecting the structure, then writing
画像・アイコン素材をキーレス優先で用意 (アイコン=lucide / 静的=ログイン済みアプリ生成→配置)。プロンプトテンプレート+共通スタイルブロックでセット品質を担保。自動生成時のみ MCP or AI Gateway(コスパ別モデル表・既定 imagen-4.0-fast)。next/image 配線。