← 返回 Skills 市场

Agents Plugin Marketplace

Name: Agents Plugin Marketplace
Author: harryhihi

作者 Harryhihi · GitHub ↗ · v1.0.0 · MIT-0

cross-platform ✓ 安全检测通过

总下载

当前安装

版本数

在 OpenClaw 中安装

/install agents-plugin-marketplace

功能描述

84 production-ready agentic plugins, 192 agents, 156 skills & 102 commands for Claude Code, Codex CLI, Cursor, OpenCode & Gemini CLI

使用说明 (SKILL.md)

Agentic Plugin Marketplace (OpenClaw)

84 production-ready agentic workflow building blocks: 84 plugins, 192 agents, 156 skills, 102 commands — built for Claude Code and consumed natively by OpenAI Codex CLI, Cursor, OpenCode, Gemini CLI, and GitHub Copilot.

Source: C:\Users\Harry\Downloads\agents\ Original: https://github.com/wshobson/agents License: MIT

本技能基於 GitHub 上的 wshobson/agents 修改與封裝。

Overview

A marketplace of agentic plugins with a single source-of-truth (plugins/ directory) mapped to multiple agent harnesses:

Harness	Status
Claude Code	Native plugin support
OpenAI Codex CLI	Supported
Cursor	Supported
OpenCode	Supported
Gemini CLI	Supported
GitHub Copilot	Supported

Structure

agents/
├── plugins/           # Plugin definitions (source of truth)
│   ├── python-development/
│   ├── web-framework/
│   ├── database/
│   └── ... (84 plugins)
├── docs/              # Documentation and harness guides
├── tools/             # Utility tools
├── AGENTS.md          # Agent definitions
└── Makefile           # Build and install commands

Usage in OpenClaw

When the user needs agentic workflow building blocks:

Browse plugins/ for available plugins
Each plugin contains agent definitions, skill references, and commands
Reference docs/ for cross-harness usage patterns

安全使用建议

Install only in a trusted ClawHub maintainer environment. Review the moderation and migration commands before use, keep admin tokens scoped, use dry runs and confirmations for writes, and consider using the autoreview helper's no-yolo mode if full-access nested review is not acceptable.

能力评估

ℹ Purpose & Capability

The skills cover ClawHub moderation, PR maintenance, Convex setup, migrations, performance review, and autoreview; high-impact capabilities such as banning users, transferring packages, sending staff email, and production migrations are directly tied to the stated repo-maintenance purposes.

✓ Instruction Scope

Sensitive actions are scoped with explicit targets, reasons, dry runs, confirmation steps, audit logging, and verification expectations; no artifact-backed prompt injection or hidden override behavior was found.

✓ Install Mechanism

The artifacts appear to be plain skill documentation plus one local autoreview helper script, with no install-time hooks, obfuscated setup, or automatic execution path observed.

ℹ Credentials

The skills expect access to repo-local tools, GitHub CLI, Convex tooling, ClawHub admin auth, and external reviewer CLIs where relevant; this is broad but disclosed and proportionate for maintainer workflows.

ℹ Persistence & Privilege

No unbounded background persistence was found; the main privilege concern is that the autoreview helper defaults nested Codex review to full-access sandbox bypass, but this is disclosed and has a no-yolo opt-out.

如何使用

确保已安装 OpenClaw（本地或 Docker 部署）
在对话框中输入安装命令：/install agents-plugin-marketplace
安装完成后，直接呼叫该 Skill 的名称或使用 /agents-plugin-marketplace 触发
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

v1.0.0

Initial OpenClaw skill: 84 agentic plugins, 192 agents, 156 skills. Based on wshobson/agents (MIT).

元数据

Slug agents-plugin-marketplace

版本 1.0.0

许可证 MIT-0

累计安装 1

当前安装数 1

历史版本数 1

常见问题