镜头设计师

镜头设计师

Created by JudeYang.

Use this skill to design production-ready image/video prompts from an approved script, client confirmation workbook, storyboard, or asset plan.

This skill owns prompt generation. Upstream producer skills such as 韦斯安德森 should call this skill instead of carrying prompt-writing rules directly.

Inputs

Use the most complete available sources:

• approved client confirmation workbook, such as 项目名_脚本与资产确认表_v01.xlsx
• internal asset production workbook, such as 设定资产制作表_v01.xlsx
• confirmed character, scene, product, prop, and effect assets under 03_设定资产/
• user-provided prompt reference or approved style sample
• picture ratio, such as 9:16竖屏 or 16:9横屏
• platform/tool target, such as Seedance, LibTV, or a generic image/video model

If required source facts are missing, mark them as 待提供 or 待确认. Do not invent product appearance, official logo details, screen text, character approvals, or client decisions.

Outputs

Typical outputs:

• internal asset image prompts inside 设定资产制作表_v01.xlsx
• 05_内部制作执行/内部执行脚本与Prompt表_v01.xlsx
• first-frame, optional middle-keyframe, and end-frame prompts
• 生成前执行说明
• 视频内容Prompt（含声音/负面）
• keyframe generation task list or visual QC notes

Python scripts may create workbook scaffolds, extract source fields, and validate finished prompts. They must not generate final prompt prose by template.

Video Prompt Output Contract

When writing 视频内容Prompt（含声音/负面）, copy this exact top-level structure. Do not rename, translate, reorder, or replace these three headings with alternatives such as 视频类型、产品一致性、声音要求、三段式结构、第一段/第二段/第三段.

【基础设定】
镜头任务：
参考素材：
人物/产品/道具/场景：
连续性：
声音：
文字策略：默认画面为无文字纯图片。禁止生成字幕、标题、角标、水印、说明文字、乱码、伪Logo、排版边框和UI界面。产品镜头仅保留官方素材中已有的品牌Logo和官方屏显信息；剧情道具文字只允许最小必要大字或后期替换占位。

【氛围与画质】
风格核心：
视觉基调：
色彩与影调：
质量边界：

【画面内容】
分镜一：00:00-00:00
景别：
机位：
构图：
运镜手法：
画面内容：
特效：
声音：
结束状态：
衔接要求：

If a field is not relevant, write a concrete negative or boundary such as 特效：无新增特效 or 参考素材：未提供实际文件，按文字设定执行；不得写成已上传参考图. Do not omit fields.