← 返回 Skills 市场
securityqq

vargai

作者 Alex · GitHub ↗ · v2.0.3 · MIT-0
cross-platform ⚠ suspicious
228
总下载
0
收藏
0
当前安装
4
版本数
在 OpenClaw 中安装
/install vargai
功能描述
Generate AI videos, images, speech, and music using varg. Use when creating videos, animations, talking characters, slideshows, product showcases, social con...
安全使用建议
This skill appears to do what it says: it uses a VARG_API_KEY to call varg.ai and supports optional local rendering with bun+ffmpeg. Before installing or running anything: 1) Inspect scripts/setup.sh and scripts/setup.ts (they are included) so you know what they modify or send; never run scripts you haven't reviewed. 2) Only supply VARG_API_KEY unless you intentionally want BYOK billing — optional provider keys (FAL_KEY, ELEVENLABS_API_KEY, OPENAI/GOOGLE keys) are sensitive and, when provided as headers, will be forwarded for single requests (the docs say header BYOK isn't stored). 3) Use preview modes (--preview or cloud preview) when iterating to avoid unexpected charges. 4) For local work, run in an isolated/dev environment if you are unsure (bun and ffmpeg required). 5) If you need higher assurance, paste the setup scripts here or inspect them manually and verify they don't exfiltrate data or fetch arbitrary code from unknown hosts.
功能分析
Type: OpenClaw Skill Name: vargai Version: 2.0.3 The varg-ai skill bundle is a legitimate integration for the varg.ai media generation service. It provides comprehensive documentation and scripts for generating AI video, images, and speech via both cloud and local rendering modes. The setup scripts (scripts/setup.sh and scripts/setup.ts) perform standard environment checks for dependencies like bun and ffmpeg and verify API key connectivity to api.varg.ai. While the skill involves high-privilege actions such as executing shell commands and sending TSX code to a remote rendering service (render.varg.ai), these behaviors are transparently documented and essential to the skill's stated purpose. No evidence of malicious intent, data exfiltration, or harmful prompt injection was found.
能力评估
Purpose & Capability
Name/description (AI video/image/speech/music via varg) align with requested artifacts: primary env var is VARG_API_KEY, anyBins lists curl or bun, and docs describe cloud (curl+API) or local (bun+ffmpeg) modes. Optional provider keys (Fal, ElevenLabs, etc.) are documented for BYOK billing and make sense for a gateway client.
Instruction Scope
SKILL.md primarily instructs how to submit TSX code to the render API, how to run local CLI (bunx vargai), and how to set up a .env with VARG_API_KEY. It references additional optional provider env vars and suggests running included setup scripts (scripts/setup.sh or bun scripts/setup.ts). The instructions do not request access to unrelated system credentials or paths, but they do encourage creating/editing a local .env and executing bundled scripts — inspect those scripts before running.
Install Mechanism
There is no formal install spec (instruction-only), which is lower risk. However two setup scripts are included (scripts/setup.sh, scripts/setup.ts) and the docs instruct running them; running bundled scripts executes code from the skill package locally. There are no external download URLs in the manifest, but you should review the scripts' contents before executing.
Credentials
The skill declares a single required credential (VARG_API_KEY) which is proportional to gateway usage. The docs additionally describe optional BYOK provider keys (FAL_KEY, ELEVENLABS_API_KEY, REPLICATE_API_TOKEN, OpenAI/Google keys for certain render-only flows). Those are optional but sensitive — only provide provider keys if you intend BYOK billing. The SKILL.md uses provider keys in request headers (header mode) and documents that header keys are not stored; saved BYOK is also mentioned (stored encrypted), which is a platform behavior to be aware of.
Persistence & Privilege
always:false and normal model invocation are used. The skill does not request system-wide persistent privileges. It may write a .env or create local cache directories per the local-render workflow (normal for CLI tools). It does not claim to modify other skills or global 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 vargai
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /vargai 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
v2.0.3
Auto-published from vargHQ/skills@35a5fc51
v2.0.2
Auto-published from vargHQ/skills@a780c576
v2.0.1
Auto-published from vargHQ/skills@87ddb962
v2.0.0
Initial publish — AI video, image, speech, and music generation skill
元数据
Slug vargai
版本 2.0.3
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 4
常见问题

vargai 是什么?

Generate AI videos, images, speech, and music using varg. Use when creating videos, animations, talking characters, slideshows, product showcases, social con... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 228 次。

如何安装 vargai?

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

vargai 是免费的吗?

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

vargai 支持哪些平台?

vargai 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 vargai?

由 Alex(@securityqq)开发并维护,当前版本 v2.0.3。

💬 留言讨论