← 返回 Skills 市场
149
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install doubao-media
功能描述
Doubao (Volcengine ARK) 文生图、文生视频工具 - 生成后自动发送到对话，无需手动查找文件。Text-to-image and text-to-video with auto-send to chat.
使用说明 (SKILL.md)
Doubao Media / 豆包媒体生成
中文: 豆包（字节跳动火山引擎ARK）文生图、文生视频工具。生成后自动发送到对话，无需手动查找文件。
English: Doubao (ByteDance Volcengine ARK) text-to-image and text-to-video tool. Auto-sends generated content to chat, no manual file search needed.
Features / 功能特点
| Feature | 中文 | English |
|---|---|---|
| Text-to-Image | ✅ 文生图 | ✅ Generate images from text |
| Text-to-Video | ✅ 文生视频 | ✅ Generate videos from text |
| Image-to-Video | ✅ 图生视频 | ✅ Generate videos from images |
| Auto-Send to Chat | ✅ 自动发送到对话 | ✅ Auto-send to conversation |
| Windows Compatible | ✅ Windows 兼容 | ✅ Windows compatible |
Prerequisites / 前置条件
Required / 必需
- ARK_API_KEY - Get from / 从控制台获取: https://console.volcengine.com/ark
- Python 3.8+ with
requestslibrary
Install Dependencies / 安装依赖
pip install requests
Set API Key / 设置 API Key
# Windows PowerShell
$env:ARK_API_KEY="your_api_key_here"
# Linux/Mac
export ARK_API_KEY="your_api_key_here"
Usage / 使用方法
Generate Image / 生成图片
python scripts/doubao_media.py img "一只可爱的橘猫在阳光下睡觉"
Result / 结果:
- Image saved to
output/directory / 图片保存到output/目录 - Auto-sent to chat / 自动发送到对话
Generate Video / 生成视频
# Sync mode (wait for completion) / 同步模式（等待完成）
python scripts/doubao_media.py vid "一只猫在草地上奔跑" --duration 5
# Async mode (return task ID) / 异步模式（返回任务ID）
python scripts/doubao_media.py vid "一只猫在草地上奔跑" --async
Generate Video from Image / 图生视频
python scripts/doubao_media.py vid "让这只猫动起来" --image "https://example.com/cat.jpg"
Check Video Status / 检查视频状态
python scripts/doubao_media.py status "task_xxxxx"
Parameters / 参数说明
Image Generation / 图片生成
| Parameter | Default | Description |
|---|---|---|
prompt |
Required | Text description / 文本描述 |
--size |
1024x1024 | Image size: 1024x1024, 1024x1536, 1536x1024 |
Video Generation / 视频生成
| Parameter | Default | Description |
|---|---|---|
prompt |
Required | Text description / 文本描述 |
--duration |
5 | Duration in seconds (2-12) / 时长（秒） |
--ratio |
16:9 | Aspect ratio: 16:9, 4:3, 1:1, 9:16 |
--image |
None | Image URL for image-to-video / 图生视频的图片URL |
--async |
False | Async mode / 异步模式 |
Available Models / 可用模型
| Model ID | Function | Description |
|---|---|---|
doubao-seedream-3-0-t2i-250415 |
Text-to-Image / 文生图 | Generate images from text |
doubao-seedance-1-0-pro-250528 |
Text-to-Video / 文生视频 | Generate videos from text or images |
Output / 输出
Generated files are saved to / 生成的文件保存到:
~/.openclaw/workspace/output/
├── img_YYYYMMDD_HHMMSS.jpeg # Images / 图片
└── vid_YYYYMMDD_HHMMSS.mp4 # Videos / 视频
Performance / 性能指标
| Operation | Time | Description |
|---|---|---|
| Text-to-Image / 文生图 | 10-30s | Depends on complexity |
| Text-to-Video / 文生视频 | 1-3min | 5-second video |
Troubleshooting / 故障排除
Error: ARK_API_KEY not set
Solution / 解决方案:
$env:ARK_API_KEY="your_api_key" # Windows
export ARK_API_KEY="your_api_key" # Linux/Mac
Error: InvalidEndpointOrModel.NotFound
Solution / 解决方案: Go to Volcengine Console → Model Management and activate the required models. 前往火山引擎控制台 → 模型管理，开通所需模型。
Required models / 需要开通的模型:
- Doubao-SeeDream-3.0-T2I (Text-to-Image)
- Doubao-Seedance-1.0-Pro (Text-to-Video)
Error: UnicodeEncodeError (Windows)
Solution / 解决方案:
The script handles this automatically with PYTHONIOENCODING=utf-8.
Security / 安全
- Never hardcode API keys / 永远不要硬编码 API Key
- Use environment variables / 使用环境变量存储敏感信息
- Regularly rotate keys / 定期轮换 API Key
References / 参考资源
Version / 版本: 1.0.0 Last Updated / 最后更新: 2026-03-29
安全使用建议
Before installing or running this skill:
- Expectation vs reality: The registry metadata incorrectly states no env vars required. You must provide ARK_API_KEY (set in environment) or the script will fail. doubao-media.json and SKILL.md correctly mention ARK_API_KEY—update or verify metadata before automated installs.
- 'Auto-send to chat' is advertised but not implemented: the script's _send_to_chat only prints a message; it does not actually call any messaging tool or send files. If you rely on auto-send, inspect/implement the messaging integration or treat sending as manual.
- Verify the API endpoint: the script calls BASE_URL = https://ark.cn-beijing.volces.com/api/v3. Confirm this is the correct, official Volcengine endpoint for your account and region.
- Review network and privacy implications: the skill downloads generated media from URLs the API returns and writes them to ~/.openclaw/workspace/output. Ensure you are comfortable with those files being saved and that your environment permits outbound connections to Volcengine and to the returned file URLs.
- Run in a sandbox first: test with a non-privileged or temporary API key and inspect network calls and saved files before using a production key.
- If you need the advertised auto-send, request the author to implement a real send (or add instructions for how OpenClaw agents should be invoked to attach/send files). Also ask the publisher to fix registry metadata to declare ARK_API_KEY as required.
功能分析
Type: OpenClaw Skill
Name: doubao-media
Version: 1.0.0
The skill is a legitimate tool for generating images and videos using the Volcengine ARK (Doubao) API. The Python script `scripts/doubao_media.py` implements standard API interaction, file downloading, and status polling without any signs of malicious behavior, data exfiltration, or unauthorized execution.
能力评估
Purpose & Capability
The description and code align with a Volcengine ARK media generator and the code legitimately requires an ARK_API_KEY. However, the top-level registry metadata declared 'Required env vars: none' while SKILL.md and doubao-media.json both indicate ARK_API_KEY is required. That mismatch is an incoherence that could mislead users or automation that relies on registry metadata.
Instruction Scope
SKILL.md instructs the user to set ARK_API_KEY and run the included Python script (consistent). It also promises 'Auto-Send to Chat' behavior; the script contains a _send_to_chat method but it only prints a message and does not actually invoke any subprocess or API to send the file. The SKILL.md therefore overstates automation: auto-send appears unimplemented (comment only). There are no instructions to read unrelated files or exfiltrate extra data.
Install Mechanism
No install spec (instruction-only) and the only runtime dependency is the 'requests' Python package (SKILL.md tells users to pip install requests). Nothing is downloaded from arbitrary URLs during install. Low install risk.
Credentials
The code correctly requires a single API key (ARK_API_KEY), which is proportionate for a cloud media-generation skill. The problem is that the registry metadata omitted this requirement, while doubao-media.json and SKILL.md declare it—this inconsistency could cause automated systems or users to miss the need for credentials. No other unrelated credentials are requested.
Persistence & Privilege
The skill does not request always:true or any elevated persistence. It writes generated files to ~/.openclaw/workspace/output (consistent with its purpose) and does not modify other skill configs or system-wide settings.
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 doubao-media - 安装完成后，直接呼叫该 Skill 的名称或使用
/doubao-media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
doubao-media 1.0.0 – Initial release
- Provides text-to-image, text-to-video, and image-to-video generation via Doubao (Volcengine ARK).
- Automatically sends generated content to chat for convenience.
- Supports both synchronous and asynchronous video generation modes.
- Compatible with Windows; requires Python 3.8+ and the requests library.
- Includes detailed usage instructions and troubleshooting tips.
元数据
常见问题
doubao-media 是什么？
Doubao (Volcengine ARK) 文生图、文生视频工具 - 生成后自动发送到对话，无需手动查找文件。Text-to-image and text-to-video with auto-send to chat. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 149 次。
如何安装 doubao-media？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install doubao-media」即可一键安装，无需额外配置。
doubao-media 是免费的吗？
是的，doubao-media 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
doubao-media 支持哪些平台？
doubao-media 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 doubao-media？
由 systiger（@systiger）开发并维护，当前版本 v1.0.0。
推荐 Skills