← Back to Skills Marketplace
44
Downloads
0
Stars
0
Active Installs
2
Versions
Install in OpenClaw
/install cinematic-storyboard-generator
Description
Generate professional cinematic storyboard prompts with seven key elements, strict asset referencing, precise lighting, material consistency, spatial coheren...
README (SKILL.md)
Cinematic Storyboard Generator
Generate professional, cinema-grade image prompts for AI generation platforms (LibTV/LibLib, 即梦, Seedance, etc.).
Quick Start
Every shot prompt must contain these 7 elements (in order):
日夜景 · 主体描述 · 空间坐标 · 镜头形式 · 光影设计 · 情绪落点 · 资产锚点
Example:
夜 · 男主立于城楼最高处，深褐短发金色发带，灰色对襟长袍金色云纹刺绣，拂尘横握 · 身后城墙延展，脚下士兵阵列，前景地面 · 远景 · 低角仰拍 · 固定镜头 · 侧逆光，金色月光从左侧45度打入，硬光，冷蓝#8899BB · 肃穆 · 战意凝聚 · 参考：宿迁_霸王故居_场景图.png + 定妆照_季沧海.png
Four Iron Rules
Rule 1: Asset Anchoring
- Character description MUST reference actual costume photo (定妆照)
- Scene description MUST reference actual scene image file
- Never imagine what an asset looks like — if you haven't seen the photo, don't write details
- Append asset paths at end of every prompt
Rule 2: Light Logic
- Single main light only — never write multiple light sources
- Direction must be specific: "侧逆光从右侧45度打入" — not "自然光"
- Contact shadows required: "投影落地面，方向与光源一致" — characters don't float
- Back/side lighting preferred — frontal soft light = AI feel = fake
Rule 3: Material Consistency
- Skin: "真实皮肤纹理，无磨皮无柔光" — not "细腻白皙"
- Clothing: "有重力垂坠感，真实褶皱，非贴图感"
- Scene-character same rendering space: "三层可信纵深：前景地面反光 → 中景人物 → 远景建筑轮廓"
Rule 4: Spatial Coherence
- Characters stand on ground with physically correct shadows
- Three-layer depth: near ground reflection → middle ground character → far ground scene outline
- Portal/doorway as spatial boundary — not VFX effect rings
- Scale reference: character height ~1/5 of frame in wide shots
Shot Sequence Design
Step 1: Analyze the Scene
- Read the script/description
- Confirm emotion curve (what does the audience feel at this moment?)
- Identify key action and turning point
Step 2: Confirm Assets First
- View all relevant costume photos (定妆照)
- View all relevant scene images
- Establish main light direction based on time of day
- Mark asset paths before writing prompts
Step 3: Build the Sequence
Write prompts one by one. Each prompt is independent and complete.
| Sequence Position | Purpose | Shot Type |
|---|---|---|
| Opening (1-3) | Establish atmosphere, introduce setting | Wide/establishing |
| Middle (4-N) | Develop tension, reveal details | Medium/close-up |
| Climax (N-2) | Emotional peak | Close-up/extreme close-up |
| Ending (last 1-2) | Release, transition | Wide/freeze frame |
Step 4: Quality Check Each Prompt
Before finishing, verify:
- Day/night marked in first clause
- Character description matches costume photo
- Scene has specific asset anchor
- Light direction is specific (no "自然光")
- Character has ground shadow (not floating)
- Skin texture without retouching
- Clothing has weight/draping
- Three-layer depth is physically coherent
- Shot type matches emotion
- Asset paths annotated at end
Scene Type Reference
Ancient/Fantasy (古装/仙侠)
Light frameworks:
| Time | Main Light | Shadow | Texture |
|---|---|---|---|
| Morning | Side backlight, golden #D4A574 | Cool blue #4A3F6B | Dust visible in light beams |
| Dusk | Side light, orange-red, low angle | Deep orange to blue-purple | Long shadows, golden rim light |
| Night | Moonlight, cool blue #8899BB + lantern warm orange #FF6B35 | Cold-warm contrast | Volumetric light, mist |
Common scenes:
- Temple/mountain: Stone steps, cloud sea, ancient bells, pine trees
- Battlefield: Yellow earth, cracked, broken banners, dust, rusty metal weapons
- Interior: Wood structure, rammed earth walls, paper windows, bronze vessels
Modern Short Drama (现代短剧)
Light: Natural interior light, window light with side backlight preferred. Avoid frontal soft light.
Scenes: Apartment, office, street, convenience store — real environments with specific details.
Overseas Format
Style: Realism, natural color grading, no stylizing filters. Skin texture realistic-level.
Frame: 9:16 vertical (short drama format) or 16:9 horizontal (film).
Shot Type Reference
| Shot | Use When | Vertical Height Ratio |
|---|---|---|
| Extreme Wide / 全景 | Establishing shot, full scene | ~1/5 of frame |
| Wide / 远景 | Scene environment, large movement | ~1/4 of frame |
| Medium / 中景 | Narrative主体, character relationships | ~1/3 of frame |
| Close-up / 近景 | Focus on performance, expressions | Chest and above |
| Extreme Close-up / 特写 | Emotion amplification, key props | Single body part |
Angle-Emotion mapping:
| Angle | Emotional Effect |
|---|---|
| Low angle upward / 低角仰拍 | Power, authority,压迫感 |
| High angle downward / 高角俯拍 | Vulnerability, disadvantage |
| Eye level / 平视 | Objective, equal, confrontational |
| Profile / 侧面 | Narrative, observational |
| Back / 背面 | Mystery, leading viewer into frame |
Transition Words
| Type | Usage | Example |
|---|---|---|
| Time | "随月光升起..." | Transition from day to night |
| Space | "画面从...过渡到..." | Scene to scene |
| Dissolve | " ↔ " | "古钟 ↔ 水面，金字浮现" |
| Hard cut | "下一秒..." | Scene switch |
| Match cut | "场景切换至..." | Cut to matching composition |
Reference Files
For detailed elements checklist, see prompt-elements.md. For scene-type-specific guidance (ancient/fantasy), see scene-ancient.md.
Usage Guidance
This skill is an authoring guide and appears internally consistent. Before installing or using it: (1) ensure the costume/scene asset files you reference actually exist and that you are comfortable sharing their filenames/paths in prompts; generated prompts append asset paths which may be sent to third-party image/video generation services — check those services' privacy policies before including private assets; (2) the skill expects the user/agent to 'view' or provide assets but does not itself fetch external files or request credentials; if you can't or won't provide concrete asset images the tool will be limited in usefulness; (3) no install or secrets are required, so the primary risk is accidental disclosure of sensitive images or filenames when copying prompts to external platforms.
Capability Analysis
Type: OpenClaw Skill
Name: cinematic-storyboard-generator
Version: 1.0.1
The skill bundle is a professional framework for generating cinematic storyboard prompts for AI image/video platforms. It contains structured instructions (SKILL.md) and reference guides (prompt-elements.md, scene-ancient.md) focused on cinematography principles like lighting logic, spatial coherence, and asset consistency. No malicious code, data exfiltration attempts, or harmful prompt injection patterns were found.
Capability Assessment
Purpose & Capability
Name/description (cinematic storyboard prompt generator) match the SKILL.md and reference files: the files provide a 7-element prompt framework, lighting/material/spatial rules, and scene references. There are no unrelated dependencies, credentials, or installs that would be unnecessary for producing guided prompts.
Instruction Scope
Instructions require the agent/operator to reference and append concrete asset file paths (costume photos and scene images) and to 'view' those assets before writing prompts. This is coherent for a prompt-authoring skill, but means the skill expects user-supplied asset files to exist and be made available to the agent. The SKILL.md does not direct reading arbitrary system files or exfiltration, but users should be aware that asset filenames/paths are placed into generated prompts and could be included when a prompt is copied to third-party generation services.
Install Mechanism
No install spec and no code files — instruction-only skill. Nothing is written to disk or downloaded, which minimizes install-time risk.
Credentials
The skill declares no required environment variables, credentials, or config paths. It does not request unrelated secrets or system access; requested inputs are limited to user-provided asset files referenced in prompts.
Persistence & Privilege
always:false and normal invocation settings. The skill does not request permanent presence, system configuration changes, or access to other skills' credentials.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install cinematic-storyboard-generator - After installation, invoke the skill by name or use
/cinematic-storyboard-generator - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.1
Enhanced description with AI video platforms coverage, trigger keywords, and cinema-grade methodology positioning
v1.0.0
Initial release: professional cinematic storyboard prompt generator with 7 elements framework, 4 iron rules, and scene-type-specific guidance for AI image/video generation.
Metadata
Frequently Asked Questions
What is Cinematic Storyboard Generator?
Generate professional cinematic storyboard prompts with seven key elements, strict asset referencing, precise lighting, material consistency, spatial coheren... It is an AI Agent Skill for Claude Code / OpenClaw, with 44 downloads so far.
How do I install Cinematic Storyboard Generator?
Run "/install cinematic-storyboard-generator" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Cinematic Storyboard Generator free?
Yes, Cinematic Storyboard Generator is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Cinematic Storyboard Generator support?
Cinematic Storyboard Generator is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Cinematic Storyboard Generator?
It is built and maintained by beermanzz (@beermanzz); the current version is v1.0.1.
More Skills