← 返回 Skills 市场

Alibaba Ai Video Wan Video

Name: Alibaba Ai Video Wan Video
Author: powersyang

作者 PowersYang · GitHub ↗ · v1.0.0 · MIT-0

cross-platform ⚠ suspicious

125

总下载

当前安装

版本数

在 OpenClaw 中安装

/install alibaba-ai-video-wan-video

功能描述

Alibaba Cloud Wanx Video Generation - Text to Video, Image to Video, Video Editing

使用说明 (SKILL.md)

阿里云万相视频生成 (Wanx Video)

专业的视频生成和处理服务，支持文生视频、图生视频、视频编辑等功能。

支持的模型

文生视频 (Text-to-Video)

模型	类型	时长	分辨率	说明
wan2.6-t2v	有声视频	5-15s	720P/1080P	⭐最新推荐
wan2.5-t2v-preview	有声视频	5-10s	480P/720P/1080P	推荐
wan2.2-t2v-plus	无声视频	5s	480P/1080P	稳定性高
wan2.1-t2v-turbo	无声视频	5s	480P/720P	快速
wan2.1-t2v-plus	无声视频	5s	720P	高质量

图生视频 (Image-to-Video)

模型	说明
wanx-i2v-first-frame	基于首帧生成视频
wanx-i2v-first-last-frame	基于首尾帧生成视频

人像动画

模型	说明
wanx-digital-human	数字人对口型 ⭐推荐
wanx-animate-anyone	舞动人像
wanx-emo	悦动人像
wanx-live-portrait	灵动人像

视频编辑

模型	说明
wanx-video-edit	通用视频编辑
wanx-video-extend	视频延展
wanx-video-style	风格重绘

快速开始

文生视频

{baseDir}/scripts/wanx-video.sh \
  "一只小猫在草地上玩耍" \
  --out /tmp/cat.mp4 \
  --model wan2.6-t2v \
  --duration 5

图生视频

{baseDir}/scripts/wanx-i2v.sh \
  /tmp/input.jpg \
  --out /tmp/output.mp4 \
  --prompt "让画面动起来"

数字人

{baseDir}/scripts/wanx-digital-human.sh \
  --image /tmp/portrait.jpg \
  --audio /tmp/voice.mp3 \
  --out /tmp/talking.mp4

参数说明

文生视频参数

--model: 模型名称（默认：wan2.1-t2v-turbo）
--duration: 视频时长（秒，默认：5）
--resolution: 分辨率（480p/720p/1080p）
--out: 输出文件路径（必填）
--negative-prompt: 负向提示词（可选）

图生视频参数

--image: 输入图片路径
--prompt: 提示词
--out: 输出文件路径

示例

示例 1：简单场景

{baseDir}/scripts/wanx-video.sh \
  "日落时分的海滩，海浪轻拍岸边，电影感" \
  --out /tmp/beach.mp4 \
  --model wan2.6-t2v \
  --duration 10 \
  --resolution 1080p

示例 2：古诗意境

{baseDir}/scripts/wanx-video.sh \
  "床前明月光，疑是地上霜。举头望明月，低头思故乡。古典中国风，水墨画" \
  --out /tmp/jingyesi.mp4

示例 3：产品宣传

{baseDir}/scripts/wanx-video.sh \
  "现代科技产品，简洁设计，白色背景，缓慢旋转展示" \
  --out /tmp/product.mp4 \
  --model wan2.2-t2v-plus

注意事项

API Key: 需要配置阿里云百炼 API Key
服务开通: 需要在百炼控制台开通视频生成服务
异步生成: 视频生成是异步的，需要等待 1-5 分钟
费用: 视频生成会产生费用
内容限制: 不能生成违规内容

故障排除

问题：Model not exist

检查模型名称是否正确
确认 API Key 已开通视频生成服务
访问 https://bailian.console.aliyun.com/ 开通服务

问题：超时

视频生成需要时间，耐心等待
检查网络连接

问题：生成质量不佳

优化提示词，增加细节描述
尝试不同的模型
添加负向提示词排除不想要的元素

相关链接

安全使用建议

Do not run these scripts in a production environment or with sensitive prompts until the credential issue is resolved. Specific actions to consider: - The scripts include a hardcoded API key (sk-...) and will use it if DASHSCOPE_API_KEY is not set — this is a major red flag. Ask the publisher to remove any embedded keys and require the user to supply their own API key. - Verify the endpoints (dashscope.aliyuncs.com) and confirm they are official for your Alibaba/Bailian account. If in doubt, consult Alibaba docs or network-monitor the calls in a sandbox. - If you or your organization accidentally used the embedded key, rotate relevant credentials immediately. - Prefer a version where the script fails if DASHSCOPE_API_KEY is missing (no fallback), and has no hardcoded secrets. Also request that the author include the missing helper scripts referenced in the README or correct the examples. - Run the code in an isolated environment (container/VM) if you want to test it, and inspect network traffic and responses before using real data or paying for generation. Given the presence of an embedded credential and minor inconsistencies, treat this skill as suspicious until the author fixes the hardcoded key and clarifies endpoint/behavior.

功能分析

Type: OpenClaw Skill Name: alibaba-ai-video-wan-video Version: 1.0.0 The skill bundle contains a hardcoded Alibaba DashScope API key (sk-96743cd524f74354867007c0207df7b8) within both scripts/generate-video.sh and scripts/wanx-video.sh, which is a major security vulnerability. While the scripts appear to perform their stated function of video generation, the inclusion of hardcoded credentials and the lack of shell argument escaping when constructing JSON payloads for curl (potential injection) pose a risk. No evidence of intentional data exfiltration or malicious backdoors was found, but the poor security practices warrant a suspicious classification.

能力评估

ℹ Purpose & Capability

Name, description, scripts and examples all target Alibaba Cloud 'Wanx' video generation and the declared env var (DASHSCOPE_API_KEY) matches the scripts' expectation. However, the repository contains an embedded API key literal (sk-...) and two slightly different video API endpoints in different scripts, which is inconsistent with a clean integration.

⚠ Instruction Scope

SKILL.md instructs running provided shell scripts which perform network calls to external endpoints (dashscope.aliyuncs.com) with the user's prompt and then download the resulting video. That is expected for this feature, but the scripts will use a built-in fallback API key if the DASHSCOPE_API_KEY environment variable is not set — meaning the skill will send data using an embedded credential without explicit user consent. The README also references additional scripts (e.g., wanx-i2v.sh, wanx-digital-human.sh) that are not present.

ℹ Install Mechanism

There is no install spec (instruction-only), which is low-risk, but the package does include executable shell scripts that will be run. No external archives or third‑party package installs are specified.

⚠ Credentials

Only one env var (DASHSCOPE_API_KEY) is declared which is proportionate for an API integration. However, both scripts contain a hardcoded API key literal and one script falls back to that literal if the env var is missing. Hardcoded credentials undermine the declared env-var model and risk unauthorized usage, credential leakage, or unexpected billing.

✓ Persistence & Privilege

Skill does not request elevated or persistent privileges (always:false), and does not modify other skills or system-wide settings. Autonomous invocation is allowed by default but not an additional privilege here.

如何使用

确保已安装 OpenClaw（本地或 Docker 部署）
在对话框中输入安装命令：/install alibaba-ai-video-wan-video
安装完成后，直接呼叫该 Skill 的名称或使用 /alibaba-ai-video-wan-video 触发
根据 Skill 的参数说明提供必要输入，即可获得结构化输出

版本历史

v1.0.0

Initial release providing access to Alibaba Cloud Wanx Video Generation: - Supports text-to-video, image-to-video, video editing, and portrait animation models. - Includes detailed usage instructions, model descriptions, and sample scripts. - Documents all configurable parameters for various generation tasks. - Lists troubleshooting steps and important usage notes (API key, service activation, costs, content guidelines). - Provides helpful resource and documentation links.

元数据

Slug alibaba-ai-video-wan-video

版本 1.0.0

许可证 MIT-0

累计安装 0

当前安装数 0

历史版本数 1

常见问题