一键导入
auto-cinematic-video-promptll
Generate Seedance-ready cinematic video prompts for text-to-video, image-to-video, start/stop frames, and storyboard workflows.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Generate Seedance-ready cinematic video prompts for text-to-video, image-to-video, start/stop frames, and storyboard workflows.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
[EN] Build vs Buy advisor. Use when users say: 'I want to build...', 'Help me create...', 'Can you code...', 'I need a tool/app/script'. Searches for existing libraries, SaaS, and open source solutions before vibe coding. Estimates token costs and provides comparison tables. [TH] ที่ปรึกษาเขียนโค้ด ใช้เมื่อผู้ใช้พูดว่า: 'อยากจะสร้าง...', 'ช่วยสร้าง...', 'สามารถเขียนโค้ดได้ไหม...', 'ฉันต้องการ...'. ค้นหาไลบรารี่, SaaS และโซลูชั่นโอเพนซอร์สที่มีอยู่แล้วก่อนเขียนใหม่
Upgrade React, Next.js, or Vite interfaces into premium, modern, responsive, accessible product UI using Tailwind CSS and shadcn/ui. Use for UI redesign, visual polish, UX review, accessibility checks, responsive QA, component states, design tokens, and production-ready frontend refactors.
Cosmetic reference storyboard prompt skill optimized for premium visual excellence, strict character identity locking, borderless contiguous grids with no white divider lines, global anti-hallucination for packaging and materials, video-friendly storyboard continuity, ambient lighting locks, exact product geometry and scale locks, full label transcription, cosmetic-specific usage intelligence, and natural hand anatomy guidelines while keeping existing schemas compatible.
Furniture reference storyboard prompt skill adapted from the original reference storyboard bundle. Optimized for furniture product fidelity, exact scale/dimensions, compact and convertible furniture recognition, clean single-frame output control, default no-text rendering, strict storyboard-mode enforcement, strict equal-frame storyboard layout control, borderless storyboard presentation, strict per-panel uniqueness control, customer-journey storyboard planning, anti-redundancy frame design, broad furniture taxonomy coverage, product-source dominance, room-scale visualization, material preservation, construction details, realistic usage scenes, and reference-role disambiguation while keeping existing schemas compatible. Includes exhaustive visual inspection, furniture taxonomy coverage, variant handling, set handling, occlusion control, product-specific QA gates, current-reference contamination rejection, mandatory person-with-product interaction coverage, floor-textile/rug product support, physical-pattern t
Generate reference-locked cinematic image prompt packages for portraits, contact sheets, angle grids, macro detail packs, and video keyframes.
Generate complete cinematic story, beat sheet, storyboard, image prompts, start/stop frame prompts, and video prompt packages.
| name | auto_cinematic_video_promptll |
| description | Generate Seedance-ready cinematic video prompts for text-to-video, image-to-video, start/stop frames, and storyboard workflows. |
This skill generates technically structured cinematic video prompts for Seedance 2-style workflows. It does not generate video directly. It produces prompts, shot plans, continuity locks, and storyboard-ready instructions for:
The skill is designed for cinematic portrait, fashion, beauty, editorial, bridal, music-video, commercial, and character continuity workflows.
The reference image or reference set is the visual ground truth. Preserve the subject identity, wardrobe, materials, colors, hair, makeup, body proportions, lighting direction, color grade, set geometry, props, and atmosphere unless the user explicitly asks to change them.
Camera movement, pose progression, framing, and shot order may change. Identity and styling continuity must not drift.
Prompts should be cinematic but practical. Use language that describes the final video clearly:
Use when there is no reference image. Build a complete cinematic scene from text.
Use one primary image. Preserve identity, wardrobe, lighting, and environment. Design camera motion and subject movement around the reference.
Use 2 to 5 images. Assign roles such as primary character, wardrobe reference, lighting reference, style reference, or environment reference.
Use a start frame as the first visual state. The video should evolve from that frame through camera movement, body movement, expression change, focus shift, or environmental reveal.
Use both start and stop frames. The video must transition plausibly from the start state to the stop state.
Create multiple shots with durations, camera positions, and action descriptions. Each shot should feel like a deliberate film shot within the same scene.
Start Frame and Stop Frame must not be the same shot unless the user explicitly requests micro-motion. In normal cinematic mode, the two frames must differ in at least two axes:
Examples of valid differences:
Invalid output:
The skill supports many presets. The most commonly useful are:
For multi-shot output:
A strong single-shot prompt should follow this structure:
Create a [duration] cinematic video in [aspect ratio]. Using the reference as the visual ground truth, preserve [identity/wardrobe/hair/makeup/environment/lighting]. The shot begins at [start visual state]. Camera [movement]. Subject [action]. The shot ends at [end visual state]. Use [lens/depth of field/lighting/color grade]. Maintain [continuity locks]. Negative prompt: [negative prompt].
A strong storyboard prompt should include:
Use precise verbs:
Avoid vague terms:
Before outputting prompts, verify: