← 返回 Skills 市场
welliu

Skills Finder

作者 wel liu · GitHub ↗ · v1.6.0 · MIT-0
cross-platform ⚠ suspicious
390
总下载
1
收藏
0
当前安装
2
版本数
在 OpenClaw 中安装
/install skills-finder
功能描述
Intelligent skill matcher that searches multiple skill marketplaces (ClawHub & Skills.sh) in real-time. Supports ANY language for user input, multi-step skil...
安全使用建议
This skill appears to do what it says (search and install skills) but has two things to consider before installing: 1) Metadata mismatch: the registry header claims no required binaries, while SKILL.md/_meta.json require Node.js/npx (and optionally curl/jq). Confirm you have Node/npm and that the registry metadata is accurate. 2) Runtime npx usage: the script runs commands like 'npx clawhub@latest' and 'npx skills find' which download and execute code from the npm registry at runtime (and use the @latest tag). That behavior is expected for a 'finder/installer' but is a supply-chain risk — a malicious or compromised npm package (or a hijacked package name) could execute arbitrary code on your system. To reduce risk, prefer pinned versions, verify the upstream package authors (clawhub, skills), inspect those CLIs' source repositories, or run in a sandboxed environment. Also verify any skill you install via this tool before granting it broader agent permissions. Optional checks: review the referenced GitHub homepage in _meta.json, ensure clawhub/skills CLIs are the official packages you expect, and consider requesting a packaged install (pinned versions or an audit) instead of runtime @latest execution.
功能分析
Type: OpenClaw Skill Name: skills-finder Version: 1.6.0 The skills-finder bundle is a utility for searching and installing agent skills from the ClawHub and Skills.sh marketplaces. The core logic in `scripts/skill-finder.sh` uses standard CLI tools (via `npx`) to perform searches and installations based on user queries, including basic language detection and task chaining logic. No evidence of data exfiltration, unauthorized access, or malicious prompt injection was found; the behavior aligns with the stated purpose of skill discovery.
能力评估
Purpose & Capability
The name/description (search and install skills from ClawHub & Skills.sh) align with the script and SKILL.md which call npx clawhub and npx skills. However registry-level 'Requirements' in the provided header claim no required binaries, while SKILL.md and _meta.json require Node.js/npx (and jq/curl). That metadata mismatch is a coherence issue and should be clarified.
Instruction Scope
SKILL.md and scripts only instruct the agent to run marketplace CLIs (npx clawhub, npx skills) and to detect language locally. They do not request unrelated files, environment variables, or external endpoints beyond the expected package tooling. Note: the script executes npx commands which will fetch and run code from remote registries at runtime — this is in-scope for 'install' behavior but expands runtime trust (see install_mechanism).
Install Mechanism
There is no packaged install spec, but the runtime uses npx to fetch and execute clawhub@latest and skills packages. Using npx @latest implicitly downloads and runs remote code each time (no pinned versions), which increases supply-chain risk. This is an expected mechanism for this purpose but is higher-risk than a pinned or audited install. Also _meta.json lists install hints that would install npm packages; the script relies on network fetches rather than local, reviewed binaries.
Credentials
The skill does not request environment variables, credentials, or config paths. The dependencies listed (Node/npx, curl, jq) are reasonable for the described behavior and there are no unexpected secret requests.
Persistence & Privilege
The skill is not always-enabled, does not declare extra privileges, and does not modify other skills or system-wide settings. It can install other skills via npx (expected for an installer/finder), but it does not itself require persistent elevated privileges.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install skills-finder
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /skills-finder 触发
  4. 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
v1.6.0
**Skills Finder 1.6.0 — Universal skill search now supports multiple marketplaces and ANY language.** - Added support for searching and installing from both ClawHub and Skills.sh (skills.sh) marketplaces. - Expanded user input detection to ALL languages and scripts—no longer limited to specific language lists. - Updated CLI options to allow choosing the skill source (ClawHub, Skills.sh, or both). - Results from both marketplaces are merged and ranked for comprehensive recommendations. - Documentation and usage examples updated to reflect universal language capabilities and new dual-source search. - Improved multi-step skill chain detection and display for complex tasks.
v1.4.0
**Skills Finder 1.4.0 introduces multi-language support and automatic skill chaining for smarter discovery.** - Added support for English, Chinese, Spanish, Japanese, and Korean, including keyword triggers and responses in the user's language. - Integrated automatic multi-step skill chain detection for complex tasks, with suggestions to bundle tasks into new composite skills. - Enabled one-click skill installation directly from search results. - Improved search filtering by skill quality and exact name matching. - Expanded auto-trigger rules to better recognize when users want to find, install, or chain skills. - Enhanced documentation with clear usage examples and command references.
元数据
Slug skills-finder
版本 1.6.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 2
常见问题

Skills Finder 是什么?

Intelligent skill matcher that searches multiple skill marketplaces (ClawHub & Skills.sh) in real-time. Supports ANY language for user input, multi-step skil... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 390 次。

如何安装 Skills Finder?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install skills-finder」即可一键安装,无需额外配置。

Skills Finder 是免费的吗?

是的,Skills Finder 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Skills Finder 支持哪些平台?

Skills Finder 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 Skills Finder?

由 wel liu(@welliu)开发并维护,当前版本 v1.6.0。

💬 留言讨论