GC Skills

Small, opinionated Codex Skills for turning rough intent into executable instructions, quiet editorial images, and restrained motion direction.

一组克制、可复用的 Codex Skills：把模糊想法变成可执行任务，把主题变成静态视觉，再把真实图片转成图生视频动态 Prompt。

Public Skills

| Skill | Start with / 输入 | Get / 输出 |
| --- | --- | --- |
| GC AI Speak Translatorgc-ai-speak-translator | A rough request that another AI needs to understand.需要交给另一个 AI、但还没说清楚的想法。 | A precise, copy-ready prompt or agent task brief.可直接复制的 Prompt 或 Agent 任务说明。 |
| GC Minimal Zine Poster v0.3.1gc-minimal-zine-poster-v0-3 | A theme, sentence, article idea, photo, content brief, or reference set.主题、句子、文章想法、照片、内容 Brief 或参考图集。 | A generated poster and final image prompt, or an evidence-based reusable visual system when analysis is requested.生成的海报与最终生图 Prompt；请求分析时，返回基于证据的可复用视觉系统。 |
| GC Still Image Motion Directorgc-still-image-motion-director | A real still image that may become an image-to-video clip.准备交给图生视频模型的真实静态图片。 | A motion, micro-motion, or static decision, stability locks, and a copy-ready image-to-video prompt.动态判断、固定项和可复制的图生视频 Prompt。 |

Choose the Right Skill

• “I know what I mean, but I do not know how to say it to AI.”

• “I have a theme or brief and want a finished editorial poster.”

• “I already have an image and need to decide how it should move.”

Use Them Independently or Together

Each Skill is a standalone repository with its own URL, stars, installation entry, documentation, and issue tracker. They do not need to be installed or used as a bundle.

They can also form one optional workflow:

text
rough human request
        ↓
gc-ai-speak-translator
        ↓
clear executable brief
        ↓
gc-minimal-zine-poster-v0-3
        ↓
poster image + final image prompt
        ↓
gc-still-image-motion-director
        ↓
motion decision + platform-ready image-to-video prompt
        ↓
image-to-video model

text
模糊的人类需求
        ↓
gc-ai-speak-translator
        ↓
清晰、可执行的任务说明
        ↓
gc-minimal-zine-poster-v0-3
        ↓
海报图片 + 最终图片 Prompt
        ↓
gc-still-image-motion-director
        ↓
动态判断 + 可交给 Seedance 等模型的图生视频 Prompt
        ↓
图生视频模型

The Motion Director writes motion direction and prompts. It does not generate the final video by itself.

Motion Director 负责动态判断和 Prompt 编写，本身不直接生成最终视频。

Shared Principles

• inspect the real input before making recommendations;
• preserve the user's intent instead of inventing context;
• make visual and execution decisions explicit;
• avoid generic effects and decorative prompt inflation;
• return outputs that can be copied into the next tool;
• state uncertainty and capability boundaries clearly.

Install

Open a Skill repository above and follow its installation instructions. Each repository is public, independently maintained, and available under the MIT License.

进入对应 Skill 仓库，按照各自 README 的安装说明操作。三个项目均为独立公开仓库，并采用 MIT License。

Collapse