YanKaFei/kitten-punch-screenwriting

kitten-punch-screenwriting,是生产级中文编剧创作与剧本诊断系统,覆盖概念超短片、短片、电影长片、剧集/连续剧与竖屏短剧/微短剧。内置主角驱动因果、场景价值转折、潜台词、中文去AI味、故事状态机、人物知识边界、关系连续性、承诺-兑现机制、14道证据式QC、单调修复与长篇记忆快照,支持从零开发、续写、改戏、对白打磨与剧本医生诊断。

Project Overview项目介绍

This is a standardized agent skill package focused on professional Chinese screenwriting, compatible with multiple AI agent platforms including Claude Code and DeepSeek Harness. To install the skill, you can clone the repository into the skills directory of your agent platform, or simply ask your agent to load the SKILL.md file directly as a writing guideline. It is built to treat story state as a maintainable ledger, where every scene is processed as a state transaction that must pass quality checks before being committed. It includes an engineering-oriented writing methodology centered on causality-first structure and evidence-based quality checks, not just a collection of loose prompts.

This skill is built for screenwriters creating Chinese-language scripts across multiple formats, from concept short films and feature films to serial dramas and vertical short dramas for social media. It guides you through building your story step by step, with 14 different quality control checks covering causality, character agency, scene structure, continuity, dialogue naturalness, and other key aspects. Only scenes that pass all checks are saved to the story ledger, and any failed checks trigger a targeted repair process that avoids overwriting entire sections of your existing work.

The entire project is released under the permissive MIT open source license, so you can use it for free for both personal and commercial projects. Before you start using the skill for the first time, you can verify the package integrity by running a simple Python script included in the repository. To run the check, navigate to the root of the kitten-punch-screenwriting directory and execute python3 scripts/validate_skill.py in your terminal. The repository also includes reference documents for different screenplay formats to help you get familiar with the workflow quickly.

这是一个面向AI代理的中文编剧创作技能包,支持Claude Code和DeepSeek Harness(DSH)等多平台代理加载使用。它定位为生产级中文编剧创作与剧本诊断系统,核心能力覆盖从零开发、续写、改戏、对白打磨、剧本诊断以及长篇故事连续性维护,内置一套“因果优先、证据循环”的工程化中文写作方法。安装方式为将整个仓库目录放入对应平台的skills目录,或直接让代理读取SKILL.md文件作为写作规范。

该技能适用于需要AI辅助创作中文剧本的创作者,支持概念超短片、普通短片、电影长片、连续剧、竖屏微短剧等多种体裁的剧本创作。典型工作流为按照SKILL.md中的规范逐步构建故事,每完成一个场景都会进行14项质量检测,只有通过检测的场景才会被提交,避免局部错误污染整体故事结构,出错后也会按来源定位修复,不会重写整体内容。

该项目采用MIT许可证开源,完全免费供个人和商业使用。用户第一次加载技能后,可以运行内置的Python校验脚本验证技能包的完整性,具体操作是进入项目根目录,执行python3 scripts/validate_skill.py完成校验。项目还提供了分体裁参考文档、运行时规范、回归测试用例等内容,帮助创作者快速上手这套结构化的编剧创作方法。

Pre-install check安装前体检Compatibility · Security兼容性 · 安全性 1 warning1 项注意
  • Only 6 stars - very few users, little community feedback星标只有 6,几乎没人在用,遇到问题缺少社区反馈
  • Not DSH-native: a multi-platform tool that may require Node / Electron or another runtime first非 DSH 原生,是多平台兼容工具:可能要先装 Node / Electron 等运行时
DSH walks through these 9 checksDSH 会逐条核对这 9 项

Compatibility兼容性

  • DSH, Node, OS and profile requirementsDSH 版本 / Node 版本 / 操作系统 / profile 是否满足要求
  • External dependencies and runtimes (Electron / Python / Docker, ...)外部依赖与运行时(Electron / Python / Docker 等)是否齐备
  • Conflicts with installed plugins: command names, skill / tool names, ports, duplicate MCP registration与已装插件是否冲突:命令名、skill / tool 重名、端口占用、重复 MCP 注册

Security安全性

  • Repo matches the facts registered here; archived or abandoned?仓库是否与页面登记一致,是否归档或长期停更
  • Safety of preinstall / install / postinstall and install.sh / setup.ps1preinstall / install / postinstall 与 install.sh、setup.ps1 是否安全
  • curl|bash, download-then-execute, obfuscation, unrelated domains → stop immediatelycurl|bash、下载即执行、混淆代码、无关域名 → 立刻停止
  • Typosquatting or unmaintained packages among the new dependencies新增依赖里有没有 typosquatting 或无人维护的包
  • Requested permissions vs. what the feature actually needs申请了哪些权限、是否超出功能所需(filesystem / network / shell / clipboard)
  • Any sudo / admin requirement, plus uninstall and rollback是否要求 sudo / 管理员权限,以及卸载与回滚方式

Anything uncertain must be marked unknown with a note on how to confirm it. This site's signal screen is a static snapshot, not a security audit.拿不准的必须标「未知」并说明要我怎么确认。本站的信号筛查是静态快照,不能替代安全审计。

Or use CLI install (for developers)或使用命令行安装(适合开发者)

CLI Install命令行安装

dsh plugin --profile web add github:YanKaFei/kitten-punch-screenwriting

把 YanKaFei/kitten-punch-screenwriting 加入你的 DSH 配置(web profile)即可启用。

READMEREADME

小猫揍人编剧强化 Skill

DeepSeek Harness Agent Skill License Version

生产级中文编剧创作与剧本诊断系统

概念超短片 | 短片 | 电影长片 | 剧集 · 连续剧 | 竖屏短剧 · 微短剧


这是什么

生产级中文编剧创作与剧本诊断系统:从零开发、续写、改戏、对白打磨、剧本医生、结构诊断与长篇连续性维护。小猫揍人编剧强化skill v1.0.0,内置一套“因果优先、证据循环”的工程化写作方法。它不是提示词合集,而是一台把“故事状态”当作账本维护的编译器:每个场景都是一次状态事务,只有通过 QC 才允许提交,失败不会污染整个故事。

核心能力

  • 主角驱动因果:高潮与转折必须来自主角已有的欲望、选择、能力与信息。
  • 场景价值转折:每场必须改变处境、关系、信息、资源、危险或选择空间之一,并留下下一场压力。
  • 故事状态机(Story SSoT):人物图谱、关系状态、物件/资源、世界规则、不可逆事件、承诺-兑现、时间线全部唯一真实可操作化。
  • 对白引擎:对白是行为而非信息朗读,每个主要人物维护自身的信息和台词样本;内置中文自然度与去 AI 味规范。
  • 承诺-兑现系统:神秘身份、特殊能力、道具、预言、被遮住的真相等制造期待的内容,BROKEN 。
  • 14 道证据式 QC:因果、主角能动性、目标、落差、行为策略、场景价值、下场压力、代价升级、知识连续、物理连续、世界规则、承诺兑现、人物声音、可拍摄性、语言自然度。
  • 单调修复:QC 失败后按来源回修对应层,禁止为修局部问题重写整体或引入新设定。
  • 长篇记忆快照:按 Sequence / 集 / 关键高潮滚动更新,保证长篇连续性。

目录结构

kitten-punch-screenwriting v1.0.0

├── SKILL.md                      # 主说明与完整工作流
├── references/
│   ├── concept-film.md           # 概念超短片
│   ├── natural-chinese.md        # 中文自然度与去 AI 味
│   └── vertical-drama.md         # 竖屏短剧
├── runtime/
│   ├── story-state.md            # 故事状态机与账本
│   ├── evidence-qc.md            # 证据式 QC 与 Gate 依赖
│   ├── promise-payoff.md         # 承诺—兑现系统
│   └── repair-protocol.md        # 失败定位与单调修复
├── tests/
│   └── regression-cases.md       # 代表性回归测试
├── scripts/
│   └── validate_skill.py         # 结构校验脚本
├── VERSION
├── MANIFEST.txt
└── SHA256SUMS

使用方式

这是一个标准 Skill 包,可作为 Claude Code / DeepSeek Harness 等 Agent 的技能加载:将本目录放入对应平台的 skills/ 目录,或直接让 Agent 读取 SKILL.md 作为写作规范。

校验完整性:

cd kitten-punch-screenwriting
python3 scripts/validate_skill.py

许可证

MIT,见 LICENSE。

← 上一个 Prev dsh-tool-todo-tree 下一个 Next dsh-web-fetch-playwright →