Manjuforge

漫剧/短剧连载创作引擎：题材或原著到剧本、分镜、配音文案、AI提示词全套产出，跨集记忆防人设崩坏与伏笔丢失。创作漫剧、短剧剧本、连载分镜、网文改编漫剧时使用。

Install

skills CLI (Claude Code, Cursor, Codex and more)

npx skills add mengxiujie/ManjuForge --skill manjuforge

Manual install

git clone --depth 1 https://github.com/mengxiujie/ManjuForge.git /tmp/ManjuForge
mkdir -p ~/.claude/skills
cp -r /tmp/ManjuForge/. ~/.claude/skills/manjuforge

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# ManjuForge 漫剧铸造 - AI漫剧连载创作引擎

## 定位

核心差异化：**别人做好看的单集，我们做活得久的IP。**

卖点顺序（对外宣传用）：知识库专业深度 > 结构化输出能力 > 短视频流量逻辑内置 > 连载记忆机制。
不要卖"记忆机制"或"X集不崩"——上下文窗口在增长，这话竞品也能说。要卖"从一句话到可发布成品的工业化流水线"。

技术壁垒诚实评估：当前是**工程壁垒**（Prompt级记忆），不是技术壁垒。宣传用 Demo 数据说话。
战略备忘录（卖点对照表/定位陷阱/长期战略/定稿文件位置）见 `references/strategy_and_positioning.md`。

## 适用场景

- 用户要从零创作一部漫剧（输入题材 → 全套产出）
- 用户要把小说/网文改编为漫剧剧本
- 用户在做连载漫剧，需要跨集保持人设和剧情连贯
- 用户需要竖屏9:16分镜（面向AI生成器）
- 用户需要批量生成漫剧内容

## 核心产出（八大模块）

每次完整创作产出以下内容：

1. **人设档案** - 角色外貌/性格/关系/台词风格/成长弧线
2. **世界观设定** - 时代/地理/规则/势力/核心矛盾
3. **剧情大纲** - 全剧主线+分集梗概
4. **开篇钩子** - 前3秒视觉锚点+悬念设计
5. **逐镜分镜** - 横屏16:9标准版 或 竖屏9:16 AI执行版
6. **情绪标注** - 每幕情绪曲线+节奏节点
7. **配音文案** - 适配短视频的台词/旁白/独白
8. **AI绘图提示词** - 每幕配套构图/光影/风格提示词

---

## 一、连载记忆机制（核心壁垒）

### 原理

用文件系统实现轻量化持久记忆。每个漫剧项目维护七个记忆文件，分两类：

**逐集记忆**（每集创作后更新）：角色档案（状态变量）、剧情摘要、伏笔表、场景布局（变更时）
**设定总纲**（初始化+定期刷新）：世界观、**故事圣经**（每5集刷新）、**角色关系图**（关系变化时）

| 文件 | 作用 | 更新时机 |
|------|------|----------|
| `角色档案.md` | 锁定人设、关系、台词风格。**区分身份不变量（锁死：外貌锚点/性格内核/台词风格）与状态变量（逐集更新：情绪/服装/伤势/关系阶段）** | 每次创作前读取；状态变量每集更新，身份不变量仅剧情重大转折时更新 |
| `世界观.md` | 锁定时代、规则、势力 | 首次创作时写入，世界观扩展时更新 |
| `场景布局.md` | 锁定场景空间结构、道具锚点、角色位置 | 首次创作时写入，场景变更时更新 |
| `剧情摘要.md` | 记录已发生的关键剧情、伏笔、时间线 | 每集创作完成后自动追加 |
| `伏笔表.md` | **Foreshadow Engine**：全剧伏笔生命周期追踪（ID/类型/状态/回收）——解决30集+"记不住伏笔" | 每集新埋伏笔立即登记；推进/回收的伏笔同步更新状态 |
| `故事圣经.md` | **Story Bible**：全剧设定唯一事实源总纲（世界规则/角色弧光/视觉母题/禁止变化）——解决"记不住设定" | 初始化创建，每5集刷新；世界观/角色变更时回写摘要 |
| `角色关系图.md` | **Character Graph**：关系结构化数据（矩阵+事件时间线）——解决"记不住关系" | 初始化创建；每集有关系的集数更新矩阵+追加事件 |

Read the full skill on GitHub

Details

Skill name: manjuforge
Repository: mengxiujie/ManjuForge
License: MIT-0
Repo updated: 13 October 2026

FAQ

What does the Manjuforge skill do?

How do I install the Manjuforge skill?

skills CLI (Claude Code, Cursor, Codex and more): npx skills add mengxiujie/ManjuForge --skill manjuforge. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Manjuforge popular?

It lives in mengxiujie/ManjuForge, which has 24 GitHub stars and 4 forks.

Related skills

Claude Fable 5 Skills

kpab/claude-fable-5-skills

10 Agent Skills built specifically for Claude Fable 5 and 5.1 — intent, boundaries, and verification hooks instead of the step-by-step micromanagement older-model skills relied on. See skills/ for the individual skill files.

AI, LLMs & AgentsIn 1 list

ClaudeEEG

Krish-mal15/ClaudeEEG

Expert EEG/MEG analysis, BCI development, and neuroscience signal processing — from raw recordings to production-grade models. Covers theory, preprocessing pipelines, ML/DL modeling, real-time classification, statistical analysis, and data ingestion.

AI, LLMs & Agents

Prompt Warrior

timoncool/prompt-warrior

Use when the user wants to analyze their communication style with AI from Claude Code logs — asks for a "prompt warrior" profile, an "AI profile", "wrapped", a psychological portrait of how they prompt, statistics of their sessions/messages, or personalized prompt-engineering recommendations based on their chat history. Works in any language (RU/EN lexicons built in).

AI, LLMs & Agents

LLM Wiki Skill

aaronoah/llm-wiki-skill

A CLI agent skill based on Karpathy's LLM Wiki — Create and maintain a persistent, interconnected Markdown knowledge base—ingesting sources, enabling queries over compiled knowledge, and ensuring consistency through linting.

AI, LLMs & Agents

Oh My Skills

AIjunja/oh-my-skills

Use when the user wants to install, launch, inspect, inventory, create, edit, archive, deduplicate, or manage local Codex, Claude, or Agent Skills with the oh-my-skills GUI or CLI.

AI, LLMs & Agents

Antirea

MockirDev/antirea

Protect a project from AI-driven reverse engineering tools like REA. Applies three complementary layers — agent instruction barriers, terminal shims, and build hardening — to make REA's output meaningless even when it runs. Use when a project owner wants maximum protection against REA-based analysis at every stage: development, CI, and distribution.

AI, LLMs & Agents

Browse the most popular Claude skills, all ai, llms & agents skills, or write your own with the skills generator.