← Back to Skills Marketplace
181
Downloads
0
Stars
0
Active Installs
4
Versions
Install in OpenClaw
/install zbj
Description
猪八戒网(ZBJ) MCP服务器 - AI智能需求发布、订单管理、服务搜索、类目匹配，连接百万服务商的一站式威客平台
Usage Guidance
This skill appears to be what it claims: a ZBJ MCP integration. Before installing: (1) create a dedicated, least-privilege API key on your ZBJ account and use that (do not reuse high-privilege or root keys); (2) confirm ZBJ_API_URL points to the official zmcp.zbj.com (or leave it unset so the default is used); changing ZBJ_API_URL to an arbitrary host would cause your API key to be sent elsewhere; (3) review and approve the npm dependency (axios) in package.json if your environment enforces dependency reviews; (4) keep sandboxing enabled and do not grant file-system or broader network permissions; (5) rotate the API key periodically and revoke it if you stop using the skill. The only metadata inconsistency is the primary credential field (registry shows none while SKILL.md marks ZBJ_API_KEY as primary) — functionally the code expects the API key, so treat ZBJ_API_KEY as the primary secret.
Capability Assessment
Purpose & Capability
Name/description, required env vars (ZBJ_API_URL, ZBJ_API_KEY), required binaries (node, npm), network permission (zmcp.zbj.com:443), and the implemented HTTP endpoints in bridge.js all align with a marketplace/MCP integration. There are no unrelated credentials or unexpected system-level requirements.
Instruction Scope
SKILL.md instructs running node bridge.js and to set ZBJ_API_KEY/ZBJ_API_URL. bridge.js only issues HTTPS requests to the configured BASE_URL, implements an internal category search, caches categories in memory, and does not read other system files or arbitrary environment variables. It does not transmit data to third-party hosts in the code provided.
Install Mechanism
No remote download/install script is embedded in the registry; package.json lists a single dependency (axios) and SKILL.md recommends npm install. This is a standard, low-risk install pattern (no obscure URLs or extraction of arbitrary archives).
Credentials
Requested env vars (ZBJ_API_URL, ZBJ_API_KEY) are expected for this integration and are used by the code. Note: registry metadata earlier listed 'primary credential: none' while the SKILL.md YAML indicates primary_credential: ZBJ_API_KEY — this is a minor metadata inconsistency but not a functional red flag. The bridge sends the API key as the X-API-Key header to the configured BASE_URL; ensure that you use a scoped key and do not point ZBJ_API_URL to an untrusted host.
Persistence & Privilege
The skill is not marked always:true, and SKILL.md indicates non-autonomous operation (requires user invocation). It requests only network access to the official host and enforces sandbox in metadata. The skill does not request filesystem or system-level privileges.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install zbj - After installation, invoke the skill by name or use
/zbj - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.3
zbj v1.0.3
- 增加自动依赖安装命令（install 字段）和 npm 支持，无需手动 npm install
- 增加 npm 作为运行时必需二进制要求
- 强化沙箱隔离及声明 autonomous: false，限制自动化运行
- 文档中新增了安全说明、API Key 使用建议和权限详细说明
- 添加故障排查与常见问题解答
- 优化环境变量和 API Key 配置方法展示
- 其他细节文本优化，提高易用性和安全性
v1.0.2
zbj 1.0.2 Changelog
- Enhanced skill description for broader use cases and improved clarity.
- Added repository URL and detailed tags/keywords to improve discoverability.
- Expanded documentation: now includes real-world enterprise/AI Agent scenarios and sample usage flows.
- Improved formatting, case studies, and getting started sections for better onboarding.
- No functional/code changes; documentation and metadata only.
v1.0.1
- 技能名称由 zbj-mcp 更改为 zbj，并同步了目录名称与使用示例中的命令前缀。
- 增加了homepage字段，指向 https://zmcp.zbj.com。
- requires.env 增加了 ZBJ_API_KEY 环境变量要求。
- 新增 primary_credential 字段，指明 ZBJ_API_KEY 为主认证凭据。
- 示例命令、安装指令及文档中的 zbj-mcp 替换为 zbj，统一技能调用方式。
- 其余功能、接口与依赖说明保持不变。
v1.0.0
Initial release of zbj-mcp, the official ZBJ MCP skill.
- Provides AI agent integration for ZBJ's one-stop service platform using Model Context Protocol (MCP).
- Implements 18 tools across demand management, order management, service/shop search, and category services.
- Includes 2 resources for configuration and category data.
- Requires Node.js and axios; minimal permissions (network access to zmcp.zbj.com only).
- Supports demand publishing, management, order processing, service/provider search, smart category matching, reviews, and secured payments.
Metadata
Frequently Asked Questions
What is 猪八戒网官方技能?
猪八戒网(ZBJ) MCP服务器 - AI智能需求发布、订单管理、服务搜索、类目匹配，连接百万服务商的一站式威客平台. It is an AI Agent Skill for Claude Code / OpenClaw, with 181 downloads so far.
How do I install 猪八戒网官方技能?
Run "/install zbj" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is 猪八戒网官方技能 free?
Yes, 猪八戒网官方技能 is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does 猪八戒网官方技能 support?
猪八戒网官方技能 is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created 猪八戒网官方技能?
It is built and maintained by nylk (@nylk); the current version is v1.0.3.
More Skills