Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

Skylv Agency Agents

v1.0.3

193 AI Expert Agents for OpenClaw. Engineering, Design, Marketing, Product, and China Market specialists. Triggers: agency agents, AI experts, expert roles,...

0· 134·0 current·1 all-time

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for sky-lv/skylv-agency-agents.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "Skylv Agency Agents" (sky-lv/skylv-agency-agents) from ClawHub.
Skill page: https://clawhub.ai/sky-lv/skylv-agency-agents
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install skylv-agency-agents

ClawHub CLI

Package manager switcher

npx clawhub@latest install skylv-agency-agents
Security Scan
VirusTotalVirusTotal
Suspicious
View report →
OpenClawOpenClaw
Benign
medium confidence
Purpose & Capability
The name/description claim a library of many expert agent roles; the SKILL.md contains role lists, triggers, usage, and orchestration examples that align with that purpose. No unrelated credentials or binary requirements are declared.
Instruction Scope
The runtime instructions are primarily guidance and examples for using the agent roles, orchestration YAML, and triggers. However, the SKILL.md explicitly instructs cloning https://github.com/jnMetaCode/agency-agents-zh.git and running ./scripts/install.sh --tool openclaw, which directs the agent/user to fetch and execute remote code outside the skill bundle. The instructions otherwise do not request unrelated files, environment variables, or system secrets.
Install Mechanism
The registry contains no install spec and the skill itself is instruction-only (no code files). Nevertheless, the instructions advise git-cloning a third-party GitHub repo and running its install.sh. Fetching and executing an external install script is a higher-risk install mechanism because the script's contents are not included in the skill package and were not vetted here.
Credentials
The skill declares no required environment variables, credentials, or config paths. The SKILL.md does not ask for secrets or unrelated system credentials. This is proportionate to the stated purpose (a prompt/role library).
Persistence & Privilege
The skill does not request always:true and does not declare modifications to other skills or system-wide settings. It is user-invocable and may be invoked autonomously by the agent platform (default), which is expected for a library of agents.
Assessment
This package is essentially a catalog of prompts/agent definitions and appears coherent with its stated purpose. The main risk is the instruction to git-clone and run an external install.sh from a third-party GitHub repo — that script could write files or execute arbitrary commands. Before running any install script: (1) manually inspect the referenced repository and the contents of scripts/install.sh (or run it in a sandbox/VM/container), (2) prefer copying only the agent definitions you need into a reviewed location instead of running an automated installer, (3) verify the upstream project and recent commit history (the SKILL.md references jnMetaCode and an upstream msitarzewski project), and (4) check licensing and any network/endpoints the install script contacts. If you cannot review the remote code, treat the install step as higher-risk and avoid running it on production or sensitive hosts.

Like a lobster shell, security has layers — review code before you run it.

latestvk97d3fh9hrt1dzxxzcf701qyxx856my5
134downloads
0stars
4versions
Updated 1w ago
v1.0.3
MIT-0

Agency Agents — 193 个 AI 专家角色库

功能说明

为 OpenClaw 引入 193 个即插即用的 AI 专家角色,覆盖工程、设计、营销、产品、游戏、安全、金融等 18 个部门。每个智能体都有独立的人设、专业流程和可交付成果。

智能体分类

🛠️ Engineering (45 个)

智能体功能触发词
Security Engineer按 OWASP Top 10 审查代码安全审查、代码审计
DevOps EngineerCI/CD、Docker、K8s 配置部署、CI/CD、容器化
Frontend ExpertReact/Vue/Angular 最佳实践前端、React、组件
Backend ExpertAPI 设计、数据库优化后端、API、数据库
ML Engineer机器学习模型训练部署机器学习、模型、训练

🎨 Design (18 个)

智能体功能触发词
UX Designer用户体验设计、用户研究UX、用户体验、交互设计
UI Designer视觉设计、设计规范UI、视觉、设计系统
Brand Strategist品牌定位、视觉识别品牌、logo、vi 设计

📈 Marketing (25 个)

智能体功能触发词
SEO SpecialistSEO 优化、关键词研究SEO、搜索排名、关键词
Content Strategist内容策略、编辑日历内容策略、选题、日历
Social Media Manager社交媒体运营、活动策划社交媒体、活动、运营

🇨🇳 China Market (46 个原创)

智能体功能触发词
小红书运营专家种草笔记、达人合作小红书、种草、笔记
抖音投放专家信息流广告、DOU+ 投放抖音、投放、信息流
微信生态专家公众号、小程序、视频号微信、公众号、小程序
B 站 UP 主顾问视频策划、粉丝运营B 站、视频、up 主
跨境电商专家亚马逊、Shopee、独立站跨境电商、亚马逊、shopify
飞书协作专家飞书多维表格、自动化飞书、协作、多维表格
钉钉开发专家钉钉小程序、机器人钉钉、小程序、机器人

🎮 Gaming (12 个)

智能体功能触发词
Game Designer游戏机制、关卡设计游戏设计、关卡、机制
Unity DeveloperUnity C# 开发、优化Unity、C#、游戏开发
Unreal DeveloperUnreal Engine 蓝图、C++Unreal、UE5、蓝图

💼 Business (22 个)

智能体功能触发词
Product Manager产品规划、需求文档产品、PRD、需求
Growth Hacker增长策略、A/B 测试增长、a/b 测试、转化
Data Analyst数据分析、可视化数据分析、图表、洞察
Financial Analyst财务分析、投资评估财务、投资、估值

使用方法

1. 安装 agency-agents

# 克隆仓库
git clone https://github.com/jnMetaCode/agency-agents-zh.git

# 安装到 OpenClaw
cd agency-agents-zh
./scripts/install.sh --tool openclaw

2. 激活智能体

用户:激活安全工程师,审查这段代码

3. 多智能体协作

用户:创建一个新功能,需要产品经理、设计师、开发工程师协作

输出:

  • 产品经理:需求分析和 PRD
  • 设计师:UI/UX 设计方案
  • 工程师:技术实现方案

智能体文件格式

每个智能体包含:

# 智能体名称

## 身份定义
- 角色:[明确的专业身份]
- 经验:[年限和背景]
- 专长:[核心技能领域]

## 关键规则
1. [必须遵守的原则]
2. [工作流程要求]
3. [交付物标准]

## 工作流程
1. [第一步:输入分析]
2. [第二步:专业处理]
3. [第三步:输出交付]

## 交付物
- [具体的输出格式和内容要求]

与普通提示词的区别

普通提示词Agency Agents
"你是一个专家"明确定义专家怎么思考、怎么做事
通用建议专业流程和可交付成果
单次使用可重复激活的持久角色
孤立工作支持多智能体协作

多智能体编排

使用 Agency Orchestrator 定义工作流:

# workflow.yaml
agents:
  - role: product-manager
    output: prd
  
  - role: ux-designer
    input: prd
    output: wireframes
  
  - role: frontend-developer
    input: [prd, wireframes]
    output: implementation

execution:
  mode: sequential # 或 parallel
  model: claude-sonnet-4
npx agency-orchestrator run workflow.yaml --input "创建一个电商首页"

支持的 AI 工具

工具支持状态安装命令
OpenClaw ⭐✅ 原生支持./install.sh --tool openclaw
Claude Code✅ 原生支持./install.sh --tool claude-code
GitHub Copilot✅ 原生支持./install.sh --tool copilot
Cursor✅ 支持./install.sh --tool cursor
Hermes Agent✅ 支持./install.sh --tool hermes
More...12 种工具见安装脚本

热门智能体 TOP 10

  1. Security Engineer - 代码安全审查
  2. Product Manager - 产品规划和 PRD
  3. UX Designer - 用户体验设计
  4. Frontend Expert - React/Vue开发
  5. DevOps Engineer - CI/CD和容器化
  6. SEO Specialist - 搜索引擎优化
  7. Content Strategist - 内容策略
  8. Data Analyst - 数据分析
  9. Growth Hacker - 增长黑客
  10. 小红书运营专家 - 小红书种草

中国市场特色智能体

平台运营类

  • 小红书运营专家
  • 抖音投放专家
  • 微信生态专家
  • B 站 UP 主顾问
  • 快手运营专家
  • 知乎内容专家

跨境电商类

  • 亚马逊运营专家
  • Shopee 运营专家
  • 独立站运营专家
  • TikTok Shop 专家

本地化类

  • 政务 ToG 专家
  • 医疗合规专家
  • 教育行业专家
  • 金融行业专家

相关文件

触发词

  • 自动:检测智能体、专家、角色相关关键词
  • 手动:/agency-agents, /experts, /agents
  • 短语:激活专家、用智能体、多智能体协作

Usage

  1. Install the skill
  2. Configure as needed
  3. Run with OpenClaw

Comments

Loading comments...