什么是 Skill Hub？

Skill Hub 是扩展 AI 智能体能力的首要门户。它作为本地环境与庞大 Openclaw 技能生态系统之间的桥梁，提供了一个统一的界面来搜索数千个精选工具。通过与 ClawHub 注册表和 GitHub 目录集成，它确保开发人员能够为其智能体工作流获取最新的扩展。

除了简单的发现功能，Skill Hub 还优先考虑生态系统的安全性。它对恶意模式、代码级漏洞和复杂的提示词注入进行深度扫描。这确保了您添加到智能体中的每一个 Openclaw 技能都是安全、可靠的，并符合技术最佳实践，为自动化工作流提供值得信赖的基础。

下载入口:https://github.com/openclaw/skills/tree/main/skills/phenixstar/skill-hub

安装与下载

1. ClawHub CLI

从源直接安装技能的最快方式。

2. 手动安装

将技能文件夹复制到以下位置之一

优先级：工作区 > 本地 > 内置

3. 提示词安装

将此提示词复制到 OpenClaw 即可自动安装。

Skill Hub 应用场景

• 当用户要求特定功能或工具集成时，寻找新的能力。
• 审计当前已安装 Openclaw 技能的安全性，以检测潜在的提示词注入风险。
• 使用命令行直接从 ClawHub 注册表安装新工具。
• 通过与最新的 Openclaw 技能仓库同步，监控生态系统的增长。

1. 枢纽查询 ClawHub 注册表和 awesome-openclaw-skills 目录，根据关键词或类别查找匹配项。
2. 它根据精选状态、成熟度和之前的审查结果等因素计算信誉评分（0-100）。
3. 用户可以触发安全扫描，分析技能逻辑是否存在危险操作，如 Shell 注入或环境收割。
4. 验证通过后，使用 clawhub CLI 工具将技能安装到本地环境中。
5. 状态仪表板跟踪已安装 Openclaw 技能相对于总可用目录的覆盖范围。

Skill Hub 配置指南

要开始管理您的 Openclaw 技能，请确保您的本地环境中具有必要的脚本和依赖项。

# 搜索可用技能
python3 scripts/skill-hub-search.py --query "spreadsheet"

# 从注册表安装特定技能
npx clawhub@latest install 

# 将本地目录与最新更新同步
python3 scripts/skill-hub-sync.py

Skill Hub 数据架构与分类体系

Skill Hub 根据细粒度的信誉系统和技术元数据对 Openclaw 技能进行分类。

安全扫描同时检查代码级风险（eval/exec）和 NLP 级风险（角色劫持、外泄提示词）。

name: skill-hub
description: "OpenClaw skill discovery, security vetting & install. Searches 3000+ curated skills from ClawHub registry and awesome-openclaw-skills catalog. Scores credibility, detects prompt injection & malicious patterns, manages installations. Quick-checks GitHub for new skills."
license: MIT
version: 1.0.0
homepage: https://github.com/PhenixStar/openclaw-skills-collection
user-invocable: true
disable-model-invocation: false
auto_activate:
  - "find skill"
  - "install skill"
  - "search skills"
  - "what skills exist"
  - "skill for"
  - "discover skill"
  - "vet skill"
  - "scan skill"
  - "skill security"
  - "new skills"
  - "skill updates"
  - "browse skills"
allowed-tools:
  - Bash
  - Read
  - Write

Skill Hub

Unified skill discovery, security vetting, and installation for OpenClaw.

Commands

Search Skills

Find skills by keyword, category, or credibility score.

python3 scripts/skill-hub-search.py --query "spreadsheet"
python3 scripts/skill-hub-search.py --category "DevOps" --min-score 60
python3 scripts/skill-hub-search.py --query "auth" --live        # include live ClawHub results
python3 scripts/skill-hub-search.py --installed                  # show only installed
python3 scripts/skill-hub-search.py --not-installed --limit 20   # discovery mode

Install Skills

After finding a skill, install via ClawHub:

npx clawhub@latest install 

Vet Skills (Security Scan)

Scan a skill for malicious patterns, prompt injection, and logic weaknesses.

python3 scripts/skill-hub-vet.py --slug google-sheets     # vet single skill
python3 scripts/skill-hub-vet.py --all-installed           # vet all installed
python3 scripts/skill-hub-vet.py --category "DevOps"       # vet category
python3 scripts/skill-hub-vet.py --top 10                  # vet top N unvetted

Status Dashboard

See installed vs catalog coverage, unvetted warnings, recommendations.

python3 scripts/skill-hub-status.py

Quick Check (GitHub API)

Fast check if new skills were added since last sync. Uses gh CLI — no full download needed.

python3 scripts/skill-hub-quick-check.py              # check for updates
python3 scripts/skill-hub-quick-check.py --sync        # auto-sync if updates found
python3 scripts/skill-hub-quick-check.py --query "ai"  # check + search new skills

Browse Full Catalog

Export catalog as formatted table (terminal or markdown), grouped by category.

python3 scripts/skill-hub-table-export.py                          # terminal table
python3 scripts/skill-hub-table-export.py --format markdown        # markdown table
python3 scripts/skill-hub-table-export.py --category "AI"          # filter category

Sync Catalog

Full re-fetch from GitHub awesome-list. Computes credibility, preserves vet results, shows diff.

python3 scripts/skill-hub-sync.py

Credibility Scores (0-100)

Security Checks

Code-level: eval/exec, shell injection, obfuscation, network access, env harvesting, destructive ops.

NLP/Prompt-level: hidden instructions, role hijacking, invisible unicode, exfiltration prompts, authority escalation, social engineering.

When to Use

• User asks "find a skill for X" or "is there a skill that can..."
• User wants to extend capabilities with new tools
• User wants to check if installed skills are safe
• Before installing unknown skills from registry