/install stoe
ZMG分镜脚本生成
Purpose
Transform an idea into a production-facing storyboard script. Optimize for clarity: directors, editors, reporters, camera operators, and AI video/image tools should understand what to create.
Inputs
Useful inputs:
- Story premise, news topic, campaign theme, or desired message.
- Duration and format: 15s, 30s, 60s, 3min, vertical, horizontal, TV package, short video, or explainer.
- Audience and platform.
- Style: documentary, warm human story, fast social edit, policy explainer, youth culture, cinematic, official, or investigative.
- Required facts, people, locations, slogans, products, or compliance constraints.
If duration is missing, assume 60-90 seconds for social video and state the assumption.
Workflow
- Restate the core story in one sentence.
- Choose a narrative structure:
- Hook -> context -> character/problem -> turning point -> resolution -> call to action.
- Problem -> evidence -> explanation -> solution -> takeaway.
- Before -> change -> after.
- Divide the runtime into scenes or beats.
- For each beat, specify image, shot type, motion, narration/dialogue, sound, text overlay, and material needs.
- Add interview questions when human subjects are involved.
- Add AI-generation prompts only when the user wants assets for image/video tools.
- Check continuity: time, place, causality, character motivation, and tone.
Output Format
Use this structure by default:
## 创作定位
- 片名/暂定标题:
- 核心表达:
- 时长与比例:
- 风格:
## 分镜脚本
| 时间 | 镜号 | 画面/场景 | 景别/运动 | 旁白/对白 | 声音 | 字幕/包装 | 备注 |
|---|---:|---|---|---|---|---|---|
## 采访设计
- 采访对象:
- 关键问题:
## 素材清单
- 实拍:
- 资料:
- AI生成:
- 包装:
## 连贯性检查
- 逻辑:
- 人物:
- 节奏:
- 风险:
Quality Rules
- Make every shot filmable or generatable.
- Avoid vague instructions like "beautiful picture"; name the visible subject, action, environment, composition, and mood.
- Keep narration speakable. Prefer short sentences.
- Do not invent factual claims for news or public-policy content.
- Flag missing source material when the script depends on unprovided facts, quotes, archive footage, or approvals.
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install stoe - After installation, invoke the skill by name or use
/stoe - Provide required inputs per the skill's parameter spec and get structured output
What is zmg-srew?
Create shootable storyboard scripts for ZMGClaw internal demos. Use when the user asks to turn a news topic, campaign concept, documentary idea, short-video... It is an AI Agent Skill for Claude Code / OpenClaw, with 52 downloads so far.
How do I install zmg-srew?
Run "/install stoe" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is zmg-srew free?
Yes, zmg-srew is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does zmg-srew support?
zmg-srew is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created zmg-srew?
It is built and maintained by jkbwfi (@jkbwfi); the current version is v1.0.0.