重要前提
安装AI Skills的关键前提是:必须科学上网,且开启TUN模式,这一点至关重要,直接决定安装能否顺利完成,在此郑重提醒三遍:科学上网,科学上网,科学上网。查看完整安装教程 →
veo-3.2-prompter by pexoai/pexo-skills
npx skills add https://github.com/pexoai/pexo-skills --skill veo-3.2-prompter此技能将用户零散的多模态素材(图像、视频、音频)和创作意图,转化为适用于 Google Veo 3.2 视频生成模型(Artemis 引擎)的结构化、可执行的提示词。它扮演着专家提示词工程师的角色,确保底层模型能输出最高质量的成果。
此技能分析所有用户输入,并生成一个单一的、经过优化的 JSON 对象,其中包含最终的提示词和推荐参数。内部工作流程(识别、映射、构建)会自动处理,不应向用户暴露。
atomic_element_mapping.md 将每个素材分类到其原子元素角色中。RawReferenceImage 系统附加参考图像。用户请求: "制作一个电影感的镜头,展示这个香水瓶在深色表面上旋转,就像奢侈品广告一样。" 用户上传 perfume.png
使用 veo-3.2-prompter 的代理:
广告位招租
在这里展示您的产品或服务
触达数万 AI 开发者,精准高效
最终输出(供内部使用):
{
"final_prompt": "Hero shot, a frosted glass perfume bottle with gold cap rotating slowly on a reflective dark surface, three-point studio lighting with soft key and rim light creating subtle caustics, smooth 180-degree arc, hyper-realistic luxury commercial style with shallow depth of field. Crystalline chime, soft ambient pad.",
"reference_images": [
{
"file": "perfume.png",
"reference_type": "SUBJECT"
}
],
"recommended_parameters": {
"model": "veo-3.2-generate",
"duration_seconds": 8,
"aspect_ratio": "16:9",
"resolution": "1080p",
"generate_audio": true
}
}
| 功能 | 能力 |
|---|---|
| 引擎 | Artemis — 世界模型物理模拟(非像素预测) |
| 最大时长 | 约 30 秒原生连续生成 |
| 音频 | 原生对话 + 同步音效 |
| 参考图像 | 最多 3 张(STYLE、SUBJECT、SUBJECT_FACE) |
| 视频扩展 | 通过前一个视频输入链接片段 |
| 首/末帧 | 指定起始和/或结束关键帧 |
| 分辨率 | 720p、1080p、4K(带超分) |
| 宽高比 | 16:9、9:16 |
此技能依赖内部知识库来做出明智的决策。代理在执行过程中必须查阅这些文件。
references/atomic_element_mapping.md : 核心知识。包含"素材类型 → 原子元素"和"原子元素 → 最佳参考方法"的映射表,已针对 Veo 3.2 的参考图像系统进行调整。references/veo_syntax_guide.md : Veo 3.2 Gemini API 语法参考,涵盖 RawReferenceImage、GenerateVideosConfig、视频扩展以及首/末帧指定。每周安装量
50
代码仓库
GitHub 星标数
388
首次出现
2026年3月9日
安装于
gemini-cli48
codex48
github-copilot47
amp47
cline47
kimi-cli47
This skill transforms a user's scattered multimodal assets (images, videos, audio) and creative intent into a structured, executable prompt for the Google Veo 3.2 video generation model (Artemis engine). It acts as an expert prompt engineer, ensuring the highest quality output from the underlying model.
This skill analyzes all user inputs and generates a single, optimized JSON object containing the final prompt and recommended parameters. The internal workflow (Recognition, Mapping, Construction) is handled automatically and should not be exposed to the user.
atomic_element_mapping.md to classify each asset into its atomic element role(s).RawReferenceImage system.User Request: "Make a cinematic shot of this perfume bottle rotating on a dark surface, like a luxury commercial." User uploadsperfume.png
Agent usingveo-3.2-prompter: The agent internally processes the request and assets, then outputs the final JSON to the next skill in the chain.
Final Output (for internal use):
{
"final_prompt": "Hero shot, a frosted glass perfume bottle with gold cap rotating slowly on a reflective dark surface, three-point studio lighting with soft key and rim light creating subtle caustics, smooth 180-degree arc, hyper-realistic luxury commercial style with shallow depth of field. Crystalline chime, soft ambient pad.",
"reference_images": [
{
"file": "perfume.png",
"reference_type": "SUBJECT"
}
],
"recommended_parameters": {
"model": "veo-3.2-generate",
"duration_seconds": 8,
"aspect_ratio": "16:9",
"resolution": "1080p",
"generate_audio": true
}
}
| Feature | Capability |
|---|---|
| Engine | Artemis — world-model physics simulation (not pixel prediction) |
| Max duration | ~30s native continuous generation |
| Audio | Native dialogue + synchronized SFX |
| Reference images | Up to 3 (STYLE, SUBJECT, SUBJECT_FACE) |
| Video extension | Chain clips via previous video input |
| First/last frame | Specify start and/or end keyframes |
| Resolutions | 720p, 1080p, 4K (with upscaling) |
| Aspect ratios | 16:9, 9:16 |
This skill relies on an internal knowledge base to make informed decisions. The agent MUST consult these files during execution.
references/atomic_element_mapping.md : Core Knowledge. Contains the "Asset Type → Atomic Element" and "Atomic Element → Optimal Reference Method" mapping tables, adapted for Veo 3.2's reference image system.references/veo_syntax_guide.md : Veo 3.2 Gemini API syntax reference, covering RawReferenceImage, GenerateVideosConfig, video extension, and first/last frame specification.Weekly Installs
50
Repository
GitHub Stars
388
First Seen
Mar 9, 2026
Installed on
gemini-cli48
codex48
github-copilot47
amp47
cline47
kimi-cli47
超能力技能使用指南:AI助手技能调用优先级与工作流程详解
56,600 周安装