lanshu-awesome-ai-video-kit
cclank/lanshu-awesome-ai-video-kit
lanshu-awesome-ai-video-kit is an agent skill project by cclank on GitHub: Toolkit from enterprise AI video projects: Seedance/Kling prompts, storyboards, model selection and debugging, with 600+ prompts and methodology write-ups. It covers storyboard and prompts, with prompts for Seedance / Jimeng and Kling. 422 stars, last updated 2026-10-07, MIT license.
Editor's notes
The repo pairs a prompt library with 7 Claude Code skills, and the README gives an ln -s loop that links each skill folder into ~/.claude/skills. seedance-storyboard splits a plot into 3 to 5 shots, each written along four dimensions: camera move, action and expression, spatial position, and audio; prompt-translator rewrites prompts using a 110-entry cross-model comparison set. The README now lists 604 prompts and 21 methodology docs, more than the 411 prompts in the repo description. Running python3 serve.py lets you browse it locally, and a GitHub Action checks 32 official endpoints every week.
- Type
- Storyboards · visual assets · video prompts
- Stages
- Storyboard · Prompts
- Video models
- Seedance / Jimeng · Kling
- Format
- Agent skill
- Language
- Chinese
- GitHub stars
- ★ 422
- Last updated
- 2026-10-07
- License
- MIT
- SKILL.md files
- 7
- Checked by hand
- 2026-10-13
In the author's words
A toolkit forced into being by enterprise AI video projects · 411 prompts · 15 models · 7 Claude Skills · 14 methodology articles
Translated from the Chinese original.
Skills in this repo (7)
Names and descriptions come from each SKILL.md frontmatter; Chinese descriptions are shown in our English translation.
- happyhorse-prompter
Generates compact prompts (30-55 words) that follow HappyHorse 1.0's strict rules: subject first + explicit camera technique + an audio activation path (with X audible / speaking English at natural pace). Can optionally add an "8s timing beat" structure. Use for triggers such as "make a video with HappyHorse", "make a 3-15 second short", "I want a video with native audio", "ASMR video prompt".
- kling-prompter
Generates video prompts that follow the rules of Kling 3.0 (Kuaishou). Three adaptive approaches: a 4-part basic formula (short videos) / a 5-layer advanced formula (story + audio) / an image-to-video-only formula (describe motion only). Kling is the cinematic model with the strongest Chinese understanding in 2026, native audio-video sync, up to 2 minutes, character-directed voices and Motion Brush. Use for triggers such as "make a video with Kling", "Kling AI prompt", "Chinese video generation", "story video with native audio", "image-to-video", "character dialogue video".
- model-selector
Based on the user's video needs (scene, duration, audio, language, platform limits, budget, whether local deployment / character consistency is needed, etc.), recommends the 1-3 best-matching models from 16 mainstream AI video models (12 commercial + 4 open-source) and explains why. Commercial coverage: Seedance 2.0, HappyHorse 1.0, Kling 3.0, Sora 2, Veo 3.1, **Gemini Omni** (new 2026-05), Runway Gen-4.5/Aleph, Pika 2.5, Hailuo 02, Hunyuan Video 1.5, Wan 2.7, Jimeng AI; open-source: LTX-Video 0.9.7, Mochi 1, CogVideoX 5B, Higg…
- prompt-translator
Converts an AI video prompt from the writing style of a source model (e.g. Sora 2) into the best-practice style of a target model (e.g. Kling 3.0 / Wan 2.7 / Veo 3.1). Based on 110 benchmark entries covering 10 scenes × 11 models (prompts/data/cross-model-matrix.json): not a rewrite by intuition but lookup-style in-context learning. 10 scenes: product / two-person dialogue / physical action / image-to-video / multi-person meeting / horror / nature time-lapse / abstract / wuxia / cute pets. Use for triggers such as "Sora is EOL, help me turn this prompt into Veo", "I have a Kling prompt and want to run it on Wan", "cross-…
- seedance-debugger
Diagnoses common problems when generating video with Seedance 2.0 (12 categories such as character ID drift / twins / subtitles / logos / style drift / jumps when extending / quality degradation / wrong effects / Chinese pronunciation / inaccurate voice timbre / noise at the end), locates the root cause and gives a fixed prompt. Use for triggers such as "my prompt isn't generating right", "weird subtitles appear in the video", "the face doesn't look like the reference image", "two identical characters appear", "the style changed", "how do I fix this prompt".
- seedance-prompter
Converts the user's natural-language video request into a prompt following the Doubao Seedance 2.0 advanced formula (8 elements: precise subject + action details + scene environment + lighting and color tone + camera movement + visual style + image quality + constraints). Use for triggers such as "write me a Seedance prompt", "generate a video prompt", "make a product ad video", "use Seedance to generate XX". If the user doesn't explicitly say Seedance but describes a single-shot complex narrative / multi-subject / cinematic video, prefer this skill as well.
- seedance-storyboard
Splits a complex plot/story outline into Seedance 2.0's "Shot 1/Shot 2/Shot 3" multi-shot structure. Each shot is organized along 4 dimensions (camera move + subject action and expression + position/spatial change + audio). Use for triggers such as "write this story as a Seedance prompt", "storyboard script", "multi-shot video", "video with a complex plot", "script to storyboard".
Install
npx skills add cclank/lanshu-awesome-ai-video-kit
# install a single skill
npx skills add cclank/lanshu-awesome-ai-video-kit --skill happyhorse-prompter
# pick agents (claude-code / codex / cursor / opencode / gemini-cli / openclaw)
npx skills add cclank/lanshu-awesome-ai-video-kit -a claude-code -a codexDefault commands use the skills CLI and were not verified one by one; follow the repo README. Read SKILL.md before installing. Install guide →
Part of these stacks
- Rendering with Seedance: You already have a script and need storyboards and prompts Seedance can use
Similar projects
- visual-skills★ 522
AI film director: dramaturgy rules such as Murch's editing principles, blocking and montage, plus exact Seedance prompt syntax
- DirectorSKILL★ 234
AI filmmaking: shot lists, keyframe and video prompts, with 20 built-in director styles such as Spielberg
- h3-storyboard-skill★ 178
MiniMax H3 shot breakdown and character performance design
Shot design for Seedance 2.5
AI fight-scene prompts: character-specific moves, cause-and-effect exchange chains, cross-shot continuity, in three versions (high intensity, slow pace, in between) for Seedance and H3