跳到正文

角色场景与制作资产管理

将剧本中的角色、场景和道具整理为一致的制作资产,记录视觉依据和后续使用要求。

开源精选视频内容27 个文件2,909 Stars浏览次数 0zenstory-ai/drama-skills MIT版本 f53fb20
开始使用

下载前阅读 SKILL.md 中的使用条件,并确认所用工具支持对应能力。

输入资料
确认后的故事或剧本 角色、场景和美术设定 阅读本节完整说明
交付内容
制作资产资料 角色场景说明与连续性依据 阅读本节完整说明
环境与依赖
只有安装、升级或排障时运行 python3 scripts/selftest.py。 阅读本节完整说明
  1. 获取完整包,保留配套文件的目录结构。
  2. 按照原文准备所需工具、依赖与输入资料。
  3. 阅读对应使用步骤与输入输出说明,在自己的环境中验证结果。

使用提示摘自原文,配套文件保留来源版本;实际运行条件以各包说明为准。

正文

获取原文件

references/stage-contract.md

本阶段只拥有 剧集/<EP>/视觉设定.md 中的人物/造型、地点/视图、道具/状态、声音身份与跨场连续性。 它继承剧本事实,不决定镜头瞬态、构图、动作终点或生成提示词。

每个判断在同一文档中写成复用、新身份、新变体或真实未决;来源用场景 ID,跨文档消费用可见标题和 稳定 ID。不要为身份决定建立 occurrence、decision、ledger 或接受记录。

本阶段规则

AST

ID Class Knowledge
AST-01 structural_invariant Before creating or reusing an asset, point to the scene ID and visible evidence that establishes it.
AST-02 reviewed_invariant Reconcile each occurrence as reuse, new identity, new variant, or unresolved—never guess an ambiguous name/pronoun.
AST-03 craft_default Separate Character/Look, Location/View, and Prop/State.
AST-04 reviewed_invariant Persistent identifying anchors and mutable state are not mixed.
AST-05 structural_invariant Every downstream reference resolves to a visible identity item and the stated current variant in 视觉设定.md.
AST-06 craft_default Track only asset facts needed for recognition, reuse, prompt writing, or continuity.
AST-07 reviewed_invariant Persistent voice identity and pronunciation refs stay separate from scene-level breath, emotion, volume, and delivery state.
AST-08 reviewed_invariant A voice reference binding states what it controls and what it must not; the take's emotion, its recording space, and its background never enter identity.
AST-09 reviewed_invariant A claim about what is audible in a reference requires a creator or rights-holder description, or an authorized listening observation bound to the inspected bytes; otherwise admission stays unverified.
AST-10 structural_invariant One accepted pronunciation of a proper noun uses one spelling throughout 视觉设定.md.
AST-11 reviewed_invariant Characters designed together are not bound to confusable references; each names the audible trait telling it apart from its nearest neighbour, and names that character.
AST-12 craft_default Selection criteria are few, audible and counter-exampled; they judge a candidate reference or a clone result, they do not stand in for one.
AST-13 structural_invariant When the project's prompt language is not the language of the entry headings, every 人物 entry a shot's 视觉依据 references reaches a stated conclusion: 画面代称 carrying the exact spelling prompt bodies use, or 画面代称:无 when no body names it. One accepted name uses one spelling, case included. 地点/道具 declare one only when a body names them.
AST-14 reviewed_invariant The era shown in 视觉设定.md and in the prompts projected from it does not contradict source facts—devices, platforms, slang, stated years. A retro or aged descriptor (旧/老式/复古/褪色/斑驳, old, worn, dusty, period, CRT) cites a source basis or an explicit creator choice.
AST-15 craft_default The project visual direction states explicit era anchors: the year or decade, and how devices, uniforms, furniture and lighting look in it. Location, costume and prop entries carry the era-specific form that image and video prompts copy.

CON

ID Class Knowledge
CON-01 structural_invariant Linked end and next start states match or have an explicit owner revision.
CON-02 reviewed_invariant Knowledge, injury, ownership, weather, light, or physical state does not teleport/regress without story cause.
CON-03 craft_default Track downstream-relevant deltas, not the whole 设定集 in every shot.
CON-04 structural_invariant A continuity change states before, after, cause/source scene, effective range, and affected visible IDs.
CON-05 taste_option Declared montage, ellipsis, dream, or subjective imagery may intentionally break ordinary continuity.
CON-06 structural_invariant A continuity change names every existing downstream document it affects; future work is described, not pre-created.
CON-07 structural_invariant A continuity lock declared in 视觉设定.md fixes one verbatim surface plus its shot and image scope; every in-scope frozen keyframe, motion prompt, and named image prompt carries that surface unchanged.

规则分级由高到低:structural_invariant(结构缺陷,阻断)、 reviewed_invariant(需证据判断)、craft_default(常用做法,可覆盖)、 taste_option(创作者选择,不作缺陷)。创作者已接受的事实优先于本表。

CONTACT HAONAN

联系haonan

医疗 AI · 产品与独立开发

刘浩南

关注医疗 AI 从 demo 到真实落地,聚焦产品化、商业化与院内工作流效率提升。

12年医疗产研经验,大湾区头部医疗AI公司前产品总监,独立开发者

医疗 AI 产品化医疗 AI 商业化院内工作流效率
刘浩南的微信联系二维码
微信扫一扫,联系 haonan