ワンクリックで
video-prompt-engineering
Optimize prompts for AI video generation platforms including Sora, Runway, Pika, and Kling
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
メニュー
Optimize prompts for AI video generation platforms including Sora, Runway, Pika, and Kling
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
Reference for querying the Atlas knowledge graph through its MCP tools — the SECONDARY enrichment/comparison layer that adds best-practice context to systems you have ALREADY scanned from your real sources (`az`, repos, dirs). Use when you need to look up nodes, edges, kinds, clusters, stats, or wiki pages in Atlas to compare against your real inventory. (atlas graph, query atlas, atlas mcp, search the graph, graph neighbors, atlas record, atlas kinds, enrichment layer)
Atlas turns your STATED NEED into a real systems atlas by SCANNING your actual sources (Azure via `az`, git repos, local dirs) and process/data mining them, THEN enriching against the Atlas knowledge graph. Use this skill when asked to inventory/map your real systems, scan your cloud + repos + directories, mine the real processes or data they contain, or collect their real constraints/gotchas. (atlas, scan my systems, inventory our azure account, map my repos, real systems atlas, process mining, data mining, collect nuances, system discovery)
This skill should be used when the user asks to "find skills in the wild", "assimilate popular workflows", "discover SKILL.md files in repos", "research external skills", "find workflow patterns", "survey the skill landscape", "what skills exist out there", or wants to investigate public repositories for extractable processes, babysitter plugins, and reusable procedural insights. Searches GitHub for SKILL.md files, classifies repos by archetype, and maintains structured research under docs/reference-repos/.
JavaScript and TypeScript documentation generation using JSDoc and TSDoc. Parse source code, generate API documentation, validate coverage, and integrate with TypeDoc for comprehensive developer documentation.
Structured debugging methodology using hypothesis-driven investigation, log analysis, and bisection to isolate and resolve defects.
Zero-knowledge circuit development using Circom and Noir languages. Supports constraint optimization, ZK-friendly cryptographic primitives, proof generation (Groth16, PLONK), and Merkle tree implementations.
SOC 職業分類に基づく
| name | video-prompt-engineering |
| id | SK-FTV-006 |
| version | 1.0.0 |
| description | Optimize prompts for AI video generation platforms including Sora, Runway, Pika, and Kling |
| specialization | film-tv-production |
| graph | {"domains":["domain:arts-culture"],"skillAreas":["skill-area:prompt-engineering","skill-area:video-processing","skill-area:media-encoding"],"roles":["role:creative-director","role:media-engineer"]} |
Create optimized prompts for AI video generation platforms that produce cinematic, production-quality footage. These prompts must communicate action, camera movement, timing, and style in a format that translates across different AI platforms.
[SCENE SETUP] + [CHARACTER/SUBJECT] + [ACTION SEQUENCE] +
[CAMERA MOVEMENT] + [LIGHTING/ATMOSPHERE] + [STYLE/AESTHETIC]
| Component | Content | Example |
|---|---|---|
| Scene Setup | Location, time, environment | "A rain-soaked Tokyo street at night" |
| Subject | Who/what appears | "A woman in a red coat" |
| Action | What happens (sequential) | "walks forward, stops, turns to look back" |
| Camera | Movement and framing | "slow tracking shot, eye level" |
| Lighting | Light sources, mood | "neon signs reflecting on wet pavement" |
| Style | Visual aesthetic | "cinematic, blade runner aesthetic" |
Strengths:
Prompt Style:
Natural language, paragraph format.
Describe the scene as if telling a story.
Include subtle details about atmosphere.
Mention specific camera movements by name.
Example:
"A close-up tracking shot follows a single snowflake as it falls
through the air, passing snow-covered pine branches and eventually
landing on a red mitten. The camera holds on the crystalline
structure of the snowflake for a beat before it begins to melt.
Soft, diffused winter light. Shallow depth of field with gentle
bokeh in the background."
Strengths:
Prompt Style:
Structured, detailed prompts.
Specify motion types explicitly.
Reference camera movements precisely.
Include duration indicators.
Example:
"Cinematic shot, slow motion. A detective in a trench coat walks
through a crowded train station. Camera dollies backward maintaining
medium shot. People blur past in the foreground. Harsh overhead
lighting creates deep shadows. 1940s noir aesthetic. 4 seconds."
Strengths:
Prompt Style:
Concise, focused prompts.
One primary action per prompt.
Strong style keywords.
Clear motion direction.
Example:
"Close-up of woman's face, wind blowing through hair,
looking off camera left, soft golden hour lighting,
cinematic film grain, subtle movement"
Strengths:
Prompt Style:
Detailed action sequences.
Step-by-step motion description.
Clear spatial relationships.
Timing indications.
Example:
"A chef in a professional kitchen. Wide shot. He tosses
vegetables in a wok, flames rise dramatically (2 sec),
plates the dish with precise movements (3 sec),
wipes his brow and smiles at camera (2 sec).
Warm kitchen lighting, steam rising, professional quality."
locked off, tripod, stable, still camera
- PAN: horizontal pivot (pan left, pan right)
- TILT: vertical pivot (tilt up, tilt down)
- DOLLY: camera moves (dolly in, dolly out, dolly alongside)
- TRACKING: follows subject (tracking shot, follow shot)
- CRANE: vertical lift (crane up, crane down)
- STEADICAM: smooth handheld (steadicam walk, floating camera)
- HANDHELD: naturalistic shake
- ZOOM: lens change (slow zoom, crash zoom)
- ORBIT: circles subject (360 orbit, arc shot)
slow, gentle, smooth, quick, whip, crash, gradual
walks → strides, shuffles, marches, stumbles
runs → sprints, jogs, dashes, bolts
looks → glances, stares, gazes, peers
turns → spins, pivots, rotates, wheels around
picks up → grabs, snatches, lifts, retrieves
slowly, gradually, suddenly, immediately,
after a beat, in one motion, over X seconds
## Scene [Number]: [Title]
### Setup
- **Location:** [Specific environment description]
- **Time:** [Time of day, lighting conditions]
- **Atmosphere:** [Weather, mood, ambience]
### Subject
- **Character(s):** [Who appears, wardrobe, positioning]
- **Key Props:** [Important objects in scene]
### Action Sequence
1. [First action with timing]
2. [Second action with timing]
3. [Third action with timing]
### Camera
- **Shot Type:** [Size and angle]
- **Movement:** [Specific movement description]
- **Speed:** [Movement speed]
### Technical
- **Duration:** [Total seconds]
- **Aspect Ratio:** [16:9, 2.39:1, etc.]
- **Style:** [Visual aesthetic reference]
### Platform Prompts
**Universal/Sora:**
[Paragraph-form natural language prompt]
**Runway:**
[Structured prompt with style keywords]
**Pika:**
[Concise action-focused prompt]
### Negative Prompt
[What to avoid: jittery motion, morphing, artifacts, etc.]
cinematic, film grain, anamorphic, 35mm film,
professional quality, movie scene, theatrical
golden hour, blue hour, harsh shadows, soft light,
rim lighting, volumetric, neon glow, practical lighting
smooth motion, fluid movement, realistic physics,
natural motion, consistent speed, seamless
atmospheric, moody, dramatic, serene, tense,
energetic, contemplative, mysterious
| Issue | Solution |
|---|---|
| Subject morphing | Describe subject consistently throughout |
| Jittery motion | Add "smooth" and "fluid" keywords |
| Wrong timing | Specify durations explicitly |
| Inconsistent style | Use strong style anchors |
| Background issues | Describe environment in detail |
| Physics problems | Describe motion realistically |