← 返回 Skills 市场
deemo-soul

Jianying Video Gen

作者 Deemo-Soul · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ⚠ suspicious
593
总下载
0
收藏
4
当前安装
1
版本数
在 OpenClaw 中安装
/install jianying-video-gen
功能描述
使用剪映(Jianying/小云雀)的 Seedance 2.0 模型自动生成AI视频。支持文生视频(T2V)、图生视频(I2V)和参考视频生成(V2V)三种模式。当用户需要生成AI视频、使用Seedance模型创作短片、或基于参考图像/视频进行风格转换时使用此技能。需要预先配置 cookies.json 登录凭证。
使用说明 (SKILL.md)

\r \r

剪映 AI 视频生成器\r

\r 通过 Playwright 自动化操作剪映(xyq.jianying.com),使用 Seedance 2.0 模型生成 AI 视频。\r \r

前置条件\r

\r

  1. Python 3.9+ + playwright 已安装\r
  2. Chromium 已通过 playwright install chromium 安装\r
  3. cookies.json — 剪映登录凭证(从浏览器导出),放在脚本同目录下\r \r
pip install playwright && playwright install chromium\r
```\r
\r
## 核心脚本\r
\r
`scripts/jianying_worker.py` — 主自动化脚本\r
\r
## 使用方式\r
\r
### 文生视频 (T2V)\r
\r
```bash\r
python3 scripts/jianying_worker.py \\r
  --cookies /path/to/cookies.json \\r
  --output-dir /path/to/output \\r
  --prompt "赛博朋克风格的长安城,飞行汽车穿梭在霓虹灯笼之间" \\r
  --duration 10s \\r
  --model "Seedance 2.0"\r
```\r
\r
### 图生视频 (I2V)\r
\r
```bash\r
python3 scripts/jianying_worker.py \\r
  --cookies /path/to/cookies.json \\r
  --output-dir /path/to/output \\r
  --ref-image /path/to/image.png \\r
  --prompt "将这张图片变成动画,镜头从左向右缓慢平移" \\r
  --duration 10s \\r
  --model "Seedance 2.0 Fast"\r
```\r
\r
### 参考视频生成 (V2V)\r
\r
```bash\r
python3 scripts/jianying_worker.py \\r
  --cookies /path/to/cookies.json \\r
  --output-dir /path/to/output \\r
  --ref-video /path/to/reference.mp4 \\r
  --prompt "画风改成宫崎骏风格,其他不变" \\r
  --duration 10s \\r
  --model "Seedance 2.0"\r
```\r
\r
### Dry-Run 模式(调试用)\r
\r
```bash\r
python3 scripts/jianying_worker.py --cookies /path/to/cookies.json --prompt "测试" --dry-run\r
```\r
\r
> 只填写表单不提交,生成 `step_*.png` 截图供检查。\r
\r
## 参数说明\r
\r
| 参数 | 默认值 | 可选值 | 说明 |\r
|------|--------|--------|------|\r
| `--prompt` | "一个美女在跳舞" | 任意文本 | 视频描述 |\r
| `--duration` | `10s` | `5s`, `10s`, `15s` | 视频时长 |\r
| `--ratio` | `横屏` | `横屏`, `竖屏`, `方屏` | 画面比例 |\r
| `--model` | `Seedance 2.0` | `Seedance 2.0`, `Seedance 2.0 Fast` | 模型选择 |\r
| `--ref-image` | 无 | 本地图片路径 | I2V 模式的参考图片 |\r
| `--ref-video` | 无 | 本地视频路径 | V2V 模式的参考视频 |\r
| `--cookies` | `cookies.json` | 文件路径 | 剪映登录凭证路径 |\r
| `--output-dir` | `.` | 目录路径 | 输出视频保存目录 |\r
| `--dry-run` | false | - | 只填表不提交 |\r
\r
## 模型与积分\r
\r
| 模型 | 积分/秒 | 5s | 10s | 15s | 特点 |\r
|------|---------|-----|------|------|------|\r
| Seedance 2.0 Fast | 3 | 15 | 30 | 45 | 快速,适合测试 |\r
| Seedance 2.0 | 5 | 25 | 50 | 75 | 高质量,正式出片 |\r
\r
## 自动化流程\r
\r
```\r
登录(cookies) → 新建 → 沉浸式短片 → 选模型 → [上传参考视频] → 选时长 → 输入Prompt → 提交\r
  → 拦截 thread_id → 导航详情页 → 轮询视频 → curl 下载 MP4\r
```\r
\r
## 提示词编写指南\r
\r
详细的提示词示例和编写技巧参见 `references/prompt-guide.md`。\r
\r
## 常见问题\r
\r
**Q: cookies 过期了怎么办?**\r
在浏览器登录 xyq.jianying.com,使用 EditThisCookie 等扩展导出 cookies.json。\r
\r
**Q: 下载 403?**\r
脚本使用 thread_id 详情页 + curl 下载,CDN 链接无需 cookie。如仍失败检查网络。\r
\r
**Q: 上传参考视频很慢?**\r
正常,8MB 视频约需 60-90 秒。脚本会自动等待最多 5 分钟。\r
安全使用建议
This skill automates the Jianying web UI and requires valid cookies.json credentials to work. Before installing: 1) Do NOT trust the included cookies.json — it contains session cookies packaged with the skill; treat that as a credential leak. Replace it with your own exported cookies if you intend to use the skill, or remove the file and provide your own. 2) Expect the skill to upload any reference image/video you point it at to the Jianying service (that is its purpose). 3) Note the odd hard-coded Windows output path (D:\SQLMessage\AI_Videos) in multiple scripts — change output paths to a safe, intended location before running. 4) Run in an isolated environment (sandbox/VM) the first time to observe behavior, and do not reuse sensitive account credentials included in the package. If you need higher assurance, ask the author for a public homepage or source repo and a version that does not include baked-in credentials.
功能分析
Type: OpenClaw Skill Name: jianying-video-gen Version: 1.0.0 The skill bundle includes a 'cookies.json' file containing active session tokens for the Jianying (xyq.jianying.com) platform, which is a significant security risk and highly irregular for a distributed skill. The scripts (mcp_server.py and jianying_worker.py) use Playwright for browser automation and subprocess.run to execute 'curl' commands with hardcoded Windows paths (e.g., 'D:\SQLMessage\AI_Videos'). While the code appears to fulfill its stated purpose of AI video generation, the inclusion of pre-authenticated credentials and the use of shell-adjacent execution for downloads are risky behaviors.
能力评估
Purpose & Capability
The name/description match the code: the scripts automate xyq.jianying.com using Playwright and require a cookies.json credential file. However, the package already contains a cookies.json with what look like real session cookies (credentials included in the repo), and multiple files hard-code a Windows path (D:\SQLMessage\AI_Videos) as a default output — those choices are not necessary for the stated purpose and are unexpected.
Instruction Scope
SKILL.md instructs the agent to automate the Jianying site, inject cookies, upload reference files, and download generated MP4s. The code follows those instructions (navigates the site, injects cookies, uploads local images/videos, fills forms, polls a task page, and downloads the MP4). It does not appear to read unrelated system files or access external endpoints beyond xyq.jianying.com. Note: by design it will upload any reference files you point it at to the remote service.
Install Mechanism
No install spec is provided (instruction+scripts only). requirements.txt lists Playwright and requests, which is consistent with the code. There is no suspicious remote download/install mechanism in the package itself.
Credentials
The skill does not request environment variables, but the repository ships a cookies.json containing session cookie values for xyq.jianying.com. Bundling credentials with the skill is disproportionate and risky: these cookies grant access to an account on the target service. Also, hard-coded default output paths (D:\SQLMessage\AI_Videos) are unusual and may cause unexpected file creation on installation/runs.
Persistence & Privilege
The skill is not marked always:true and does not request elevated platform privileges. It runs subprocesses to invoke local scripts and creates output directories, which is expected. There is no indication it modifies other skills or system-wide agent settings.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install jianying-video-gen
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /jianying-video-gen 触发
  4. 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 of jianying-video-gen. - Supports AI video generation using Jianying's Seedance 2.0 model. - Allows text-to-video (T2V), image-to-video (I2V), and reference video-to-video (V2V) workflows. - Automated with Playwright; requires Python 3.9+, Chromium, and a Jianying cookies.json file. - Flexible parameters including prompt, duration, ratio, model, reference image/video, and dry-run mode. - Provides usage examples, parameter guide, model points cost table, and troubleshooting FAQ.
元数据
Slug jianying-video-gen
版本 1.0.0
许可证 MIT-0
累计安装 4
当前安装数 4
历史版本数 1
常见问题

Jianying Video Gen 是什么?

使用剪映(Jianying/小云雀)的 Seedance 2.0 模型自动生成AI视频。支持文生视频(T2V)、图生视频(I2V)和参考视频生成(V2V)三种模式。当用户需要生成AI视频、使用Seedance模型创作短片、或基于参考图像/视频进行风格转换时使用此技能。需要预先配置 cookies.json 登录凭证。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 593 次。

如何安装 Jianying Video Gen?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install jianying-video-gen」即可一键安装,无需额外配置。

Jianying Video Gen 是免费的吗?

是的,Jianying Video Gen 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Jianying Video Gen 支持哪些平台?

Jianying Video Gen 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 Jianying Video Gen?

由 Deemo-Soul(@deemo-soul)开发并维护,当前版本 v1.0.0。

💬 留言讨论