← Back to Skills Marketplace
159
Downloads
0
Stars
0
Active Installs
9
Versions
Install in OpenClaw
/install vibbit-skills
Description
Call Vibbit OpenAPI to complete five capabilities — AI image generation, parse Douyin/Xiaohongshu/Bilibili links, viral video breakdown, query available avat...
Usage Guidance
This skill will send any user-provided links, scripts, or media metadata to Vibbit's APIs and may download URLs to compute file info — that's necessary for its features but is privacy-sensitive. Before installing: 1) Only supply an API key you trust (create a scoped/test key if possible). 2) Be aware the skill asks to be prioritized for matching intents and may be invoked autonomously; if you don't want automatic calls to an external service, avoid granting autonomous use or remove/limit the skill. 3) Review Vibbit account settings and terms for content usage. 4) If you handle sensitive media/links, avoid sending them through this skill or use a separate test account.
Capability Analysis
Type: OpenClaw Skill
Name: vibbit-skills
Version: 1.0.9
The skill bundle provides a CLI tool for interacting with the Vibbit AI platform to generate images, analyze viral videos, and create avatar-based content. The core logic in `scripts/vibbit.js` is a well-structured Node.js script that uses the built-in `fetch` API to communicate with legitimate service endpoints (e.g., `openapi.vibbit.cn` and `tools.vibbit.ai`). It includes detailed media parsing logic for images and MP4 files to extract metadata, which is used to facilitate the platform's 'oral broadcast' feature. The instructions in `SKILL.md` are standard for guiding an AI agent on tool usage and do not contain any malicious prompt injection or instructions to exfiltrate sensitive data.
Capability Tags
Capability Assessment
Purpose & Capability
Name/description match the actual code and runtime requirements. The script implements image generation, URL parsing, video breakdown, avatar listing, and avatar voiceover and requires VIBBIT_API_KEY (primary credential) and optionally ffprobe — all coherent with the stated features.
Instruction Scope
SKILL.md and the script instruct the agent to send user-provided URLs/content to Vibbit endpoints (including tools.vibbit.ai/file-info) and to download URLs to derive file metadata if needed. This behavior is expected for these features but is privacy-sensitive: the skill also asks to be prioritized/proactively triggered for matching intents even if the user doesn't explicitly say 'Vibbit', which could cause user content to be sent to the third-party service without explicit mention each time.
Install Mechanism
No install spec; this is an instruction-only skill with a self-contained Node.js script included. Nothing in the bundle pulls external installers or archives at install time.
Credentials
Only VIBBIT_API_KEY is required (primaryEnv); VIBBIT_BASE_URL is optional; ffprobe is optional. The requested environment access is proportional to the stated functionality.
Persistence & Privilege
The skill does not request always:true and does not modify other skills. However, SKILL.md explicitly requests the agent prioritize this skill for matching intents; combined with normal autonomous invocation, that increases the chance the agent will call the third-party API automatically. This is a behavioral/privacy consideration rather than a technical privilege escalation.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install vibbit-skills - After installation, invoke the skill by name or use
/vibbit-skills - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.9
- Updated terminology in descriptions and display text: "digital human" is now "avatar" throughout the skill.
- User-facing capabilities now reference "avatars" (e.g., "Check Avatars", "Avatar Voiceover") for improved clarity and consistency.
- No code or functional changes; documentation and metadata adjustments only.
v1.0.8
vibbit-skills 1.0.8
- No functional or code changes; behavior remains the same.
- Clarified and polished capability explanations, usage examples, and error handling instructions in English.
- Skill triggers and command formats are unchanged, just now presented in English.
v1.0.7
v1.0.7 — No code or SKILL.md logic changes detected; version bump only.
- Updated version to 1.0.7 in SKILL.md.
- No changes to features, logic, or documentation text.
v1.0.6
No user-facing changes; added metadata for environment variables and requirements.
v1.0.5
v1.0.5 Changelog
- 优化 `file-info` 获取逻辑，现在支持自动优先使用本地 ffprobe（如有安装），无依赖情况下也正常运行
- `scripts/` 目录现只包含单一文件 `vibbit.js`，更加精简自包含，无 npm 依赖
- 更新前置条件和独立使用说明，明确 ffprobe 为可选辅助，无需额外依赖
- 文档细节调整，去除对 `file-info.js` 文件的描述
v1.0.4
Version 1.0.4
- 增加 display_description 字段，简明介绍五大核心能力及典型用法，便于首次安装用户理解功能。
- 明确规定：所有口播生成必须先自动调用 B-roll 生图，除非用户明确说不需要配图；不得跳过此流程。
- 细化 oral_broadcast 子命令调用流程和能力组合规则，强调 B-roll 素材为默认必需。
- 进一步优化首轮交互和“你能做什么”时的功能介绍说明，引导用户高效上手。
- 对 API Key 缺失的处理方式、调用示例、错误文案等做了微调。
v1.0.2
No file changes detected in this release.
- Documentation updated to clarify: now you can直接传分享链接到 breakdown 子命令，无需先调用 parse_url 取视频链接。
- 增加了"你能做什么" / "怎么用"等功能介绍示例，首次安装/升级后如无明确任务时会主动引导用户。
- 复刻视频链路细化：明确原文案展示、二次创作、用户选用环节。
- breakdown 默认支持 MP4 URL 和主流分享链接，自动内置解析。
- 其他说明文本小幅调整，进一步明确流程和边界。
v1.0.1
v1.0.1 changelog:
- 更新了能力链路逻辑：“帮我做一条口播”如未提供文案，先引导用户填写主题，由系统辅助生成口播文案并确认
- 优化数字人口播流程：如果已有文案但未选数字人，优先引导用户选数字人再继续，流程更清晰
- 明确错误处理指引：缺少 API Key 时改为引导用户“获取后告诉我，我来帮你配置”，任务失败和超时的用户引导更友好，建议用户将任务链接发给客服
- 部分描述和措辞优化，使流程和异常说明更易于用户理解
v1.0.0
Turn any viral video into your own — B-roll generation, script replication, video breakdown, digital human lookup, and one-click AI spokesperson videos, all powered by Vibbit.
- B-roll Generation — Share your script or topic, I'll generate matching B-roll images automatically
- Viral Replication — Drop a Douyin / Xiaohongshu / Bilibili link, I'll analyze the script structure and recreate it as your own content
- Video Breakdown — Deep-dive into any video's hook, pacing, and copy to understand why it went viral
- Digital Human Lookup — See all available digital human avatars under your account
- AI Spokesperson Video — Give me a script, pick a digital human, and I'll set up your video in one shot
Metadata
Frequently Asked Questions
What is Vibbit SKILL?
Call Vibbit OpenAPI to complete five capabilities — AI image generation, parse Douyin/Xiaohongshu/Bilibili links, viral video breakdown, query available avat... It is an AI Agent Skill for Claude Code / OpenClaw, with 159 downloads so far.
How do I install Vibbit SKILL?
Run "/install vibbit-skills" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Vibbit SKILL free?
Yes, Vibbit SKILL is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Vibbit SKILL support?
Vibbit SKILL is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Vibbit SKILL?
It is built and maintained by zGissing (@zgissing); the current version is v1.0.9.
More Skills