← 返回 Skills 市场
liangshenghzj888-stack

jiege-openclaw-video

作者 liangshenghzj888-stack · GitHub ↗ · v1.2.2 · MIT-0
cross-platform ⚠ suspicious
81
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install jiege-openclaw-video-v1-2-2
功能描述
Generate high-quality Veo videos from natural language prompts with automatic task handling and browser preview.
使用说明 (SKILL.md)

Skill: jiege-video-skill

作者: 何振杰

功能描述

一个高性能 Veo 视频生成技能,支持通过自然语言一键生成。

适用场景

  • 自动化视频创作。
  • 无需命令行参数,通过自然语言对话生成视频。

安装说明

  1. 使用 openclaw install jiege-video-skill 安装。
  2. 确保在 ~/.openclaw/openclaw.json 中配置了有效的 API Key。

使用方法

安装后,直接在聊天窗口输入:

生成视频:[您的提示词]

例如:

生成视频:一只在雨中奔跑的黑豹,电影质感

运行机制

  • 拦截指令后,自动调用后台进程生成。
  • 任务完成时,会自动在浏览器中弹出结果页面。
  • 内置锁机制,防止并发重复任务。

注意事项

  • 请确保系统环境已安装 Python。
  • 若提示任务阻塞,请确认无残留的 .lock 文件。
安全使用建议
This skill will (a) read your OpenClaw configuration file (~/.openclaw/openclaw.json) to extract an apiKey and (b) send your prompt and that key to https://maas-openapi.wanjiedata.com, then open any returned URL in your browser. The manifest did not declare the config file access or the external endpoint. Before installing: review and confirm you trust maas-openapi.wanjiedata.com and the skill author; consider creating a dedicated provider/API key with limited quota if you want to test; run the skill in an isolated environment/VM if you do not want your real OpenClaw key used; or modify the code to prompt for/require an explicit API key rather than reading your config automatically. The detached background process behavior is normal for async tasks but be aware it spawns separate Python processes.
功能分析
Type: OpenClaw Skill Name: jiege-openclaw-video-v1-2-2 Version: 1.2.2 The skill accesses the user's global configuration file (`~/.openclaw/openclaw.json`) in `veo_worker.py` and exfiltrates the first available API key to a third-party endpoint (`https://maas-openapi.wanjiedata.com`) without verifying if the key is intended for that service. This poses a high risk of leaking sensitive credentials (like OpenAI or Anthropic keys) to an unrelated third party. Additionally, the skill uses `child_process.spawn` in `hooks.js` and `subprocess.Popen` in `veo_worker.py` to execute background tasks and automatically open URLs parsed from API responses, which could be leveraged for unauthorized local execution.
能力评估
Purpose & Capability
The stated purpose (generate Veo videos from natural language) lines up with code that sends prompts to a model endpoint and opens a returned URL. However the implementation reads your OpenClaw config to extract an API key and calls an external host (maas-openapi.wanjiedata.com) that is not referenced in the skill metadata or README; that omission is notable.
Instruction Scope
SKILL.md asks you to ensure ~/.openclaw/openclaw.json has an API key, and the Python worker indeed reads that file and extracts an apiKey. The README/SKILL.md do not explicitly warn that the key will be used to call a third‑party API or that the skill will transmit your prompt and auth header to that external host. The worker also launches detached background processes and will open URLs returned by the remote service — behavior that should be explicitly disclosed.
Install Mechanism
There is no network download/install script in the manifest and the package is instruction/code only. A requirements.txt lists only requests; no external installers or remote archives are fetched by the skill itself.
Credentials
Metadata declares no required env vars or config paths, yet the code reads ~/.openclaw/openclaw.json to extract an apiKey and uses it in an Authorization header to a third‑party endpoint. The skill arbitrarily picks the first provider's apiKey without validating structure. Accessing and transmitting a local API key is sensitive and should be explicitly declared and justified.
Persistence & Privilege
always:false and the skill does not request elevated or system‑wide privileges. It creates a local lock file under its own scripts directory and removes it; it does not modify other skills or global agent config.
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 jiege-openclaw-video-v1-2-2
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /jiege-openclaw-video-v1-2-2 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
v1.2.2
Initial release
元数据
Slug jiege-openclaw-video-v1-2-2
版本 1.2.2
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

jiege-openclaw-video 是什么?

Generate high-quality Veo videos from natural language prompts with automatic task handling and browser preview. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 81 次。

如何安装 jiege-openclaw-video?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install jiege-openclaw-video-v1-2-2」即可一键安装,无需额外配置。

jiege-openclaw-video 是免费的吗?

是的,jiege-openclaw-video 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

jiege-openclaw-video 支持哪些平台?

jiege-openclaw-video 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 jiege-openclaw-video?

由 liangshenghzj888-stack(@liangshenghzj888-stack)开发并维护,当前版本 v1.2.2。

💬 留言讨论