專為 Suno 與 Udio 設計的 AI 音樂提示詞（Prompt/Tags）生成助手。當使用者提出抽象的情緒、畫面或曲風融合需求（如：「古風結合賽博龐克」）時，請觸發此技能。它會將模糊的靈感精準轉譯為專業的音樂流派、樂器、混音與結構標籤，大幅提升 AI 音樂生成的品質與準確度。
日本語の概要は準備中です。原文の説明を表示しています。
Kouzui-HUANG/biz-ai-video-course42026年10月27日 更新
無料GitHub で公開
seedance-2-0
Executable tool skill that ACTUALLY CALLS the GMI Cloud `seedance-2-0-260128` API to generate real videos — it runs a bundled Python script (submit → poll → download), it is NOT a prompt writer. Activation is by MODEL NAME ONLY. Trigger EXCLUSIVELY when the user explicitly names the model: "seedance 2", "seedance2", "seedance 2.0", "seedance-2-0", "seedance-2-0-260128", or "seedance 2.0 生成/出片". CRITICAL — DO NOT trigger on generic video-related wording: "影片生成", "文生影", "圖生影", "動態生成", "動態提示詞", "video generation", "text-to-video", "image-to-video", "i2v", "t2v", etc. — those describe VIDEO PROMPT tasks handled by prompt-master-text-to-video / prompt-master-image-to-video / prompt-master-video-continuity and are UNRELATED to this skill. Also do NOT trigger on bare "seedance" WITHOUT "2" (Seedance 1.x / generic prompt-target mentions belong to the prompt skills) nor on "seedance 2 fast" / "seedance 2.0 fast" (that is a different model, seedance-2-0-fast-260128). Only the explicit model name above activates this skill; if the user merely describes a video or asks for a prompt but never names this model, do NOT use this skill.
日本語の概要は準備中です。原文の説明を表示しています。
Kouzui-HUANG/biz-ai-video-courseスター 4(GitHub 4 + cccskills 0)
実行スコア 未検証
インストール方法を見る含まれるファイル（4）
- SKILL.md7.6 KB
- docs/SKILL_zh-TW.md5.6 KB
- generate_video.py21.8 KB
- references/gmi-api.md5.7 KB
SKILL.md（原文）
インストールする前に、エージェントに与えられる指示の中身を確認できます。
Seedance 2.0 — Live Video Generation
This is an executable tool skill, not a prompt generator. It runs a bundled Python program that calls the real GMI Cloud seedance-2-0-260128 API, then downloads the produced video(s) and thumbnail(s). By default each run produces two variants (--count 2) at 720p.
⚠️ Activation guard — model name ONLY
Use this skill only when the user explicitly names the model — seedance 2 / seedance2 / seedance 2.0 / seedance-2-0[-260128].
This project contains MANY skills that write video prompts. Generic wording — 影片生成 / 文生影 / 圖生影 / 動態生成 / 動態提示詞 / video prompt / t2v / i2v — does NOT belong here; let the relevant prompt-writing skill handle those (a bare "Seedance" mention as a prompt target also belongs there). If no explicit model name from the list appears, do not activate. seedance 2 fast / seedance 2.0 fast is a different model (seedance-2-0-fast-260128) — also do not activate.
Files
In this skill folder:
| Path | Purpose | Git |
|---|---|---|
generate_video.py | The program (stdlib only, Python 3.8+) | tracked |
references/gmi-api.md | Full API spec — read when debugging | tracked |
.gmi_api_key | API key (Bearer; same GMI key as gemini-3-pro-image) | gitignored |
Shared media folders at the project root — one drop-zone and one results folder for all three executable generation skills (gemini-3-pro-image, gpt-image-2, seedance-2-0), wired up by .claude/lib/media_paths.py:
| Path | Purpose | Git |
|---|---|---|
uploads/ | Drop local reference images/videos/audio here (shared) | gitignored |
uploads/done/ | Materials move here after a successful run | gitignored |
outputs/ | Videos + thumbnails land here, prefixed seedance_ (shared) | gitignored |
Workflow
- Confirm activation — a model name from the list is present. If not, stop.
- Identify the task (mode is chosen automatically from the materials given):
- Text-to-video (T2V) — only a prompt.
- Reference-to-video (R2V) — prompt + image(s). This is the DEFAULT whenever the user provides/uploads images without saying otherwise: pass them as
--ref/--ref-file(→reference_images, subject/style reference). - First/last-frame (I2V) — only when the user explicitly asks for 首幀/尾幀/首尾幀 anchoring (the video starts from / ends on the exact image): use
--first-frame(-file)/--last-frame(-file). - Reference video / audio / GMI asset IDs are also supported:
--ref-video(-file),--ref-audio(-file),--ref-asset-id.
- Gather inputs:
prompt(required). If the user gives a rough idea, you may compose a strong English motion/camera prompt for them.- Materials (optional): public URL →
--ref <url>etc.; local file →--ref-file <path>etc. The script auto-uploads local files to a temp host (tmpfiles.org, ~1 h auto-delete) and uses the resulting URL. Bare filenames are looked up in the shareduploads/, thenuploads/done/. Guards: images ≤ 10 MB (PNG/JPEG/WebP/HEIC/HEIF, ≤ 9), videos ≤ 50 MB, audio ≤ 20 MB. (GMI accepts public URLs only — no base64/local upload to the API itself.)
- Parameters — apply defaults; ask the user only for a genuinely undecided necessary value:
--count(default 2) — unless the user explicitly asks for a specific number, keep the default: every run delivers TWO variants to choose from. Each video is a separate API request submitted in parallel (max 4 — video is expensive).--resolution480p/720p/1080p(default 720p)--duration4–15 seconds (default 5)--ratio16:9/4:3/1:1/3:4/9:16/21:9/adaptive— when omitted the script picksadaptiveif any visual reference is given (follows the source material), else16:9.- Audio is generated by default; pass
--no-audioto disable.--seed,--watermark,--web-searchexist for special cases.
- Run it — CWD doesn't matter, the script finds the key and the shared folders on its own. Video takes minutes, so set a generous Bash timeout (≥ 10 min):
For long/complex prompts, write the prompt to a temp file and passS=.claude/skills/seedance-2-0/generate_video.py # text-to-video python3 $S --prompt "A thunderstorm rolls over a medieval castle at night, lightning illuminating stone towers" # reference-to-video with a local image (bare filename → shared uploads/; DEFAULT mode for user images) python3 $S --ref-file character.png \ --prompt "The character walks through a neon-lit alley, cinematic tracking shot" # first/last-frame anchoring (only when explicitly requested) python3 $S --first-frame-file bud.jpg --last-frame-file bloom.jpg \ --prompt "The flower slowly blooms, petals unfurling in morning light" # single video only (when the user explicitly asks for one) python3 $S --count 1 --prompt "..."--prompt "$(cat /tmp/p.txt)"to avoid shell-quoting issues. Use--dry-runto preview the payload without spending quota. - Show the results — the script prints each saved
.mp4path and a_thumb.jpgper video under the sharedoutputs/. Open each thumbnail with the Read tool so the user gets a visual preview, and give the full.mp4paths (videos can't be previewed inline — the user opens them in Finder/QuickTime). - Used materials self-archive — on success the script moves every local reference it used (images, videos, audio, first/last frames) out of
uploads/intouploads/done/, so the drop-zone only ever shows what still needs doing. Pass--keep-refswhen the user wants to try several prompts on the same material. Re-runs still accept the bare filename — the lookup coversdone/too.
Notes
- API key precedence:
--api-key>$GMI_API_KEY>.gmi_api_key>../gemini-3-pro-image/.gmi_api_key(same GMI Cloud key for the whole platform). The key file is already present and gitignored. - Privacy — local-file flags upload to a public temp host (~1 h then auto-deleted). For sensitive material, ask the user for a private URL and use the URL flags instead.
- Cost — video is billed per generated second; keep
--count/--durationmodest unless asked. - Full request/response schema, status values, and limits:
references/gmi-api.md.
レビュー
まだレビューはありません。使ってみた感想をお寄せください。
コメント
ログインなしで書き込めます。使ってみた感想、質問、動かなかった環境などを気軽にどうぞ。
まだコメントはありません。最初のコメントを書いてみませんか。
同じリポジトリのスキル
AI Short-Drama Director & Clip Prompt Architect (AI 短劇導演 × 分鏡 Clip 架構師). Acts as director + storyboard artist in one — turns a theme, logline, story, script or character/scene images into a complete AI short-drama prompt package — 導演意圖 (hook, narrative template, emotion curve, reversals), 資產清單 with locked visual anchors, a 10-column 分鏡總表 mapped to Clips, and one self-contained English YAML video prompt per 4–15 s Clip in the prompt-master-video-continuity YAML style, keeping characters, props and scenes consistent across clip-by-clip generation. Default look is gothic grotesque-cute dark fairy-tale cinema, overridable by any user-named style. Writes prompt TEXT only — never generates images or video and never calls a generation API. Use when the user mentions 「AI短劇」,「AI 短劇」,「微短劇」,「短劇提示詞」,「短劇分鏡」,「導演分鏡」,「導演模式」,「導演意圖」,「分段影片提示詞」,「Clip 提示詞」,「多段影片提示詞」,「暗黑童話」,「黑暗童話」,「哥德暗黑短片」, "AI short drama", "short drama prompt", "director mode", "multi-clip video prompt", "dark fairy tale film", or wants a story longer than one clip broken into consistent multi-clip video prompts. Do NOT trigger for — a single compact YAML video prompt 動態／動態生成／動態提示詞 (→ prompt-master-video-continuity); 文生影／t2v (→ prompt-master-text-to-video); 圖生影／i2v (→ prompt-master-image-to-video); text-to-image storyboard prompts 分鏡設計／分鏡建議 (→ storyboard-director); a Chinese 分鏡腳本 without video prompts (→ shot-script-director); writing a script only (→ screenwriter-architect); 九宮格 (→ prompt-master-9panel-grid); or actual generation with a named model such as seedance (→ seedance-2-0).
日本語の概要は準備中です。原文の説明を表示しています。
Kouzui-HUANG/biz-ai-video-course42026年10月27日 更新
Blocking Plan Director (運鏡平面圖導演). Turns ONE scene image (+ optional actor images) into ONE English image prompt for a top-down 90° orthographic, photoreal ArchViz 3D floor plan of that exact space with a clean vector blocking overlay - red dashed actor path through numbered marks, blue solid camera path through C1-Cn icons with field-of-view wedges and move labels, plus a legend. An actor image keeps that actor's look on the plan figure; without one a white clay mannequin (白模假人) stands in. Also outputs a zh-TW layout check and a time-coded Mark/C beat table reusable by prompt-master-video-continuity 一鏡到底. Generates immediately, prompt text only. Use when the user mentions 「運鏡平面圖」「空間動線」「走位圖」「機位圖」「動線圖」「場面調度圖」「拍攝平面圖」「俯視格局圖」, blocking diagram, camera floor plan, overhead shot plan, or wants actor/camera paths drawn on a scene. Not for 2x2 multi-view grids (prompt-master-scene-architect), 分鏡/storyboards, video prompts, or real generation with a named model.
日本語の概要は準備中です。原文の説明を表示しています。
Kouzui-HUANG/biz-ai-video-course42026年10月27日 更新
Hollywood Screenwriter & Character Architect (Oscar-grade, with a sociologist's and psychologist's acuity). Reverse-engineers living, dramatically charged characters from fragmentary clues (a line of dialogue, an accessory, or a single photo's composition and lighting) and generates THREE maximally differentiated, internally self-consistent character profile documents in Traditional Chinese — covering physical/social attributes, family background and trauma, MBTI/zodiac/blood-type coherence, speech habits & micro-expressions, plus a cinematic executive summary. Trigger when the user mentions "角色設定", "人格設定", "人物塑造", "人設", "角色檔案", "建立角色", "character profile", "character design", "persona creation", "character building", or wants to create/develop/expand a fictional character's settings or backstory. Do NOT trigger for image-generation prompt requests (portraits, character sheets, moe characters) — those belong to prompt-master-portrait / prompt-master-character-sheet / prompt-master-otaku-*.
日本語の概要は準備中です。原文の説明を表示しています。
Kouzui-HUANG/biz-ai-video-course42026年10月27日 更新
Chief Character Costume & Visual Design Director. Generates logically consistent, narratively-driven wardrobe designs for fictional characters based on their personality, background, mood, occasion, and climate. Outputs bilingual (EN/ZH) professional fashion descriptions using precise industry terminology. Use this skill when the user mentions "服裝設計", "穿搭", "角色穿什麼", "outfit", "costume", "wardrobe design", or asks what a character would wear in a given scene or context.
日本語の概要は準備中です。原文の説明を表示しています。
Kouzui-HUANG/biz-ai-video-course42026年10月27日 更新
Cultural Resonance Naming Master (文化共鳴命名大師) — a screenwriting-grade naming & renaming consultant that maps a character (or any narrative entity — faction, place, weapon, project, story title) onto a global mythological archetype and forges a name carrying metaphor, foreshadowing, and historical weight. Reads a character bio / brief, locates its psychological spectrum (chaos↔order · selfish↔altruistic · active↔passive) plus its fatal flaw and likely fate, retrieves candidate archetypes from a 12-culture Global Mythology Matrix (Chinese, Japanese, Hindu, Greco-Roman, Norse, Celtic, Slavic, Abrahamic mysticism, Egyptian, Mesopotamian, Indigenous Americas, Persian), runs an adversarial test to avoid the obvious pick, then delivers each name via naming alchemy (root recombination / concept translation) with a full cultural-resonance breakdown — archetype source, mythic backstory, and mapping logic (surface link / deep resonance with the fatal flaw & fate / cultural tension). Output analysis is in Traditional Chinese. Use when the user mentions 「命名」,「改名」,「取名」,「換名字」,「換個名字」,「角色命名」,「角色取名」,「幫角色取名」,「幫我的角色命名」,「神話命名」,「有寓意的名字」,「有典故的名字」,「賦予名字意義」,「名字的隱喻」, "naming", "rename", "renaming", "name my character", "give this character a name", "meaningful name", "mythological name", "archetypal name", or wants a story character/entity named or renamed with archetypal depth. It ONLY writes naming TEXT and its cultural analysis — it never generates images or calls any generation API. Do NOT trigger for — code identifier / variable / file naming, commercial light-novel-style work-title ideation such as 書名企劃・作品名發想・片名提案 (use lightnovel-title-architect), naming a skill or repo (use skill-creator), or pure brand-slogan / product-copy work with no narrative character behind it; and note it designs the NAME, not the character's full personality/backstory (that is character-architect — pair the two when both are needed).
日本語の概要は準備中です。原文の説明を表示しています。
Kouzui-HUANG/biz-ai-video-course42026年10月27日 更新