從標準旁白與最終配音生成完整的、帶時間軸的中文字幕。適用於需要全面字幕覆蓋、可讀分段、關鍵詞強調、輸出 SRT 或 JSON，以及自動檢查字幕是否遺漏旁白的垂直知識影片。
此技能專為 AI 知識影片的字幕生成與校驗而設計。它強制使用最終確定的旁白文字作為字幕來源，並以通過審核的配音音檔作為時間軸依據，確保字幕與語音精準同步。輸出的格式包括包含時間軸資訊的 JSON 與 SRT 檔案，以及 HTML 字幕元素，支援一次顯示一句 cue，每句 cue 依中文閱讀習慣設定為 7–14 個字，必要時可擴展至 16–18 個字以保持語義完整。此外，技能還提供關鍵詞的克制性強調功能，不改變原文。最後，技能內建自動化驗證流程，通過執行 `scripts/verify-caption-coverage.ps1` 比對標準旁白與最終 HTML 的字幕覆蓋率，一旦發現遺漏即阻止交付，確保產出品質。適用於 Claude Code、Cursor 等支援自訂技能的工具平台。
npx skills add https://github.com/Vange-wang/ai-knowledge-video-skills/tree/HEAD/skills/ai-video-captions使用最終確定的旁白作為字幕文字來源，並以通過審核的配音作為時間軸依據。切勿以摘要取代完整字幕。
產出字幕 JSON、SRT 檔案、旁白到 cue 的對照表，以及合成所需的 HTML 字幕元素。一次顯示一句 cue。目標每句 cue 包含 7–14 個中文字，僅在語義單位不可分割時允許 16–18 個字。在保持文字不變的前提下，使用克制性的關鍵詞強調。
針對標準旁白與最終 HTML 執行 scripts/verify-caption-coverage.ps1。若比對結果不一致，則阻止交付。如果旁白或配音時間軸發生變更，請重新產生時間軸並再次執行驗證器。
This repo contains 10 individual skills — each has its own dedicated page.
Coordinate the full production of a vertical AI knowledge video from mixed source materials through analysis, script, visuals, mandatory character and voice approval, captions, HyperFrames composition, motion, calibration, and delivery. Use when the user wants an end-to-end short-video workflow or needs to resume or revise one stage without losing downstream dependencies.
Perform the final global calibration and delivery gate for a HyperFrames vertical video. Use to inspect all visual positions and animation paths, eliminate accidental overlap or drift, verify full captions and technical encoding, render the final MP4, and export a storyboard for review.
Complete the mandatory character-design gate for a vertical explainer video. Use when the specific presenter character is not yet chosen, when generating candidates and action sheets, or when replacing an approved character while preserving consistent proportions and clean transparent pose assets.
Analyze mixed source materials for a vertical AI knowledge video, including images, outlines, copy, scripts, videos and benchmarks. Use before writing the script to separate user instructions from source content and extract the audience problem, core argument, evidence, structure and visualization opportunities.
Assemble a portrait knowledge-video composition in HyperFrames from approved script, character, voice, captions and visual assets. Use to build deterministic 1080x1920 scene structure, media integration, safe-area layout and a paused GSAP timeline before detailed motion and sound polish.
Add controlled motion, scene transitions, background music and sound effects to an assembled HyperFrames vertical video. Use when the static composition is approved and needs tight pacing, readable animation, clean transitions and voice-first audio mixing.
Write a tight Chinese short-video script for AI knowledge and business education from an approved content analysis. Use to produce the 3–5 second hook, narrative progression, full narration, rhythm, visual notes, audio cues and CTA for a portrait video.
Produce and normalize illustrations, diagrams, icons, cards, backgrounds and other visual assets for a 9:16 AI knowledge video. Use after the visual plan to build a coherent asset library with clean hierarchy, consistent style and no accidental text or alignment defects.
Turn an approved narration into a detailed 9:16 storyboard and shot plan. Use to decide scene timing, character actions, illustrations, diagrams, background treatment, captions, transitions and sound cues before asset production or HyperFrames implementation.
Complete the mandatory voice-selection and voiceover gate for a vertical knowledge video. Use when the concrete voice is unresolved, when comparing providers or voices, when generating an approval preview, or when producing final narration with preset or authorized cloned speech.
Extract frames or short clips from videos using ffmpeg.
Summarize or transcribe URLs, YouTube/videos, podcasts, articles, transcripts, PDFs, and local files.
Use when testing, fixing, or extending the OpenClaw Control UI GUI with Vitest + Playwright end-to-end checks, mocked Gateway WebSocket flows, mocked dashboard runs, screenshots/videos, or agent-verifiable browser proof.
Obra superpowers agent skills.
See, Understand, Act on video and audio. See- ingest from local files, URLs, RTSP/live feeds, or live record desktop; return realtime context and playable stream links. Understand- extract frames, build visual/semantic/temporal indexes, and search moments with timestamps and auto-clips. Act- transcode and normalize (codec, fps, resolution, aspect ratio), perform timeline edits (subtitles, text/image overlays, branding, audio overlays, dubbing, translation), generate media assets (image, audio...
End-to-end marketing campaign planning and execution. Covers audience research, positioning, campaign angle definition, landing page copy, email sequences, social posts, ad copy, short-form video scripts, and content calendars. Use as the orchestration layer for multi-channel product launches. Use when planning or executing a multi-channel product launch, or producing landing page, email, social, or ad copy.