菜单
【导演模式·结构化描述 / Director · Structured Captioning】触发词:结构化描述 / HoloCine / 全局+角色+分镜 / 省token / 角色标签 / [char1] / 锁定外观 / structured caption / token saving。Use when 多格提示词反复重描角色外貌、token 浪费、外观微漂移时:改 HoloCine 结构 GLOBAL → CHARACTER DEFINITIONS → PER-SHOT(只用 [char1] 标签引用,不重描外貌),省约 40% token 并锁定唯一外观。
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
基于 SOC 职业分类
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。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/2799662352/ai-image-master --skill structured-captioning命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
正在显示 SKILL.md
Use when the user is on the CATIMATION 视频工作台 (video workbench) page, or asks to 批量出片 / 排成卡片 / 一次落板 / 铺满工作台 / 跑一批镜头; when a board of shot cards needs reordering, regenerating or version comparison; or when a 剧本 / 分镜表 / shot list / 制片包 should become a board. Load the catimation-video entry together with this one — this is a leaf under that entry, never a replacement for it. Symptoms: 该拆的镜头挤在一张卡里超了时长, 跨卡漂脸漂服装漂风格, 建完卡直接开跑, 拿批次摘要当验片.
Use for any question about a codebase, its architecture, file relationships, or project content — especially when graphify-out/ exists, where the question should be treated as a graphify query first. Turns any input (code, docs, papers, images, videos) into a persistent knowledge graph with god nodes, community detection, and query/path/explain tools.
Model-agnostic 12-item completeness and media-profile leaf for complex video prompt work. Load with whichever prompt base the task's model selects (2.0, 2.5, or 万相 3.0) for path B, multishot, mixed-media, or expanded director / reference work; path A may skip it. Provides format-free content definitions, live-action/2D/3D profiles, language defaults, and checks. No routing or QA.
| name | structured-captioning |
| description | 【导演模式·结构化描述 / Director · Structured Captioning】触发词:结构化描述 / HoloCine / 全局+角色+分镜 / 省token / 角色标签 / [char1] / 锁定外观 / structured caption / token saving。Use when 多格提示词反复重描角色外貌、token 浪费、外观微漂移时:改 HoloCine 结构 GLOBAL → CHARACTER DEFINITIONS → PER-SHOT(只用 [char1] 标签引用,不重描外貌),省约 40% token 并锁定唯一外观。 |
| appliesTo | ["designAndAssemble"] |
| priority | 2 |
把提示词拆成 GLOBAL(场景/时间/天气/情绪)→ CHARACTER DEFINITIONS([char1]/[char2] 锚点)→ PER-SHOT(只用标签引用,不重描外貌),省 ~40% token 并锁定唯一外观。
director-anchor-extraction-quality)。STRUCTURED CAPTIONING — separate global context from per-shot details (HoloCine pattern):
Prompt Structure (strict order):
Why This Matters:
Per-Shot Format: Panel N: [shot type], [char1] does X while [char2] does Y, [lighting], [composition]
Anti-Patterns:
GLOBAL: dim cyberpunk bar, night, rain outside, smoky neon mood. CHARACTERS: [char1]: 30s woman, silver bob, scar on left cheek, black trench. [char2]: old bartender, bald, grey beard, maroon vest.
| 错误 | 正确 |
|---|---|
| 每格重描"长发红衣女" | 只用 [char1] 标签引用 |
| 角色定义混进分镜 | 定义集中在 CHARACTER 段 |
| 省掉 GLOBAL 段 | 先写全局锚定整段 |