openskills MCP — 使用场景, 安装 & 实时演示

为什么要用

核心特性

Rust runtime — fast, predictable
TS and Python bindings for embedding into any agent
Progressive disclosure: metadata → body → scripts only when needed
Sandboxing via Seatbelt (macOS) and experimental WASM
Cross-agent: not tied to Claude Code

实时演示

实际使用效果

openskills-skill.replay ▶ 就绪

0/0

安装

选择你的客户端

~/Library/Application Support/Claude/claude_desktop_config.json · Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "openskills-skill": {
      "command": "git",
      "args": [
        "clone",
        "https://github.com/Geeksfino/openskills",
        "~/.claude/skills/openskills"
      ],
      "_inferred": true
    }
  }
}

打开 Claude Desktop → Settings → Developer → Edit Config。保存后重启应用。

~/.cursor/mcp.json · .cursor/mcp.json

{
  "mcpServers": {
    "openskills-skill": {
      "command": "git",
      "args": [
        "clone",
        "https://github.com/Geeksfino/openskills",
        "~/.claude/skills/openskills"
      ],
      "_inferred": true
    }
  }
}

Cursor 使用与 Claude Desktop 相同的 mcpServers 格式。项目级配置优先于全局。

VS Code → Cline → MCP Servers → Edit

{
  "mcpServers": {
    "openskills-skill": {
      "command": "git",
      "args": [
        "clone",
        "https://github.com/Geeksfino/openskills",
        "~/.claude/skills/openskills"
      ],
      "_inferred": true
    }
  }
}

点击 Cline 侧栏中的 MCP Servers 图标，然后选 "Edit Configuration"。

~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "openskills-skill": {
      "command": "git",
      "args": [
        "clone",
        "https://github.com/Geeksfino/openskills",
        "~/.claude/skills/openskills"
      ],
      "_inferred": true
    }
  }
}

格式与 Claude Desktop 相同。重启 Windsurf 生效。

~/.continue/config.json

{
  "mcpServers": [
    {
      "name": "openskills-skill",
      "command": "git",
      "args": [
        "clone",
        "https://github.com/Geeksfino/openskills",
        "~/.claude/skills/openskills"
      ]
    }
  ]
}

Continue 使用服务器对象数组，而非映射。

~/.config/zed/settings.json

{
  "context_servers": {
    "openskills-skill": {
      "command": {
        "path": "git",
        "args": [
          "clone",
          "https://github.com/Geeksfino/openskills",
          "~/.claude/skills/openskills"
        ]
      }
    }
  }
}

加入 context_servers。Zed 保存后热重载。

claude mcp add openskills-skill -- git clone https://github.com/Geeksfino/openskills ~/.claude/skills/openskills

一行命令搞定。用 claude mcp list 验证，claude mcp remove 卸载。

使用场景

实战用法： openskills

How to embed Agent Skills into your own agent runtime

👤 Engineers building custom LLM agents ⏱ ~120 min advanced

何时使用： You're writing an agent from scratch and want a skill system that already understands SKILL.md.

前置条件

Rust toolchain (for building) or prebuilt binding — rustup install stable; or use the TS/Python binding
Repo cloned — git clone https://github.com/Geeksfino/openskills ~/.claude/skills/openskills

步骤

Install the binding

Install the openskills Python or TS binding and point it at my skills directory.✓ 已复制

→ Runtime loaded; skills enumerated
Expose to your agent

Wire the runtime into my agent's planner so it considers skills at each step.✓ 已复制

→ Agent has a SkillRegistry abstraction
Run sandboxed

Enable Seatbelt sandboxing for skill scripts on macOS.✓ 已复制

→ Skill scripts run under sandbox profile

结果： Your custom agent can load and run SKILL.md bundles like Claude Code does.

注意事项

Loading all scripts at startup kills memory — Rely on progressive disclosure — metadata first
Sandbox denies legitimate operations — Audit the sandbox profile; narrow-then-widen rather than default-deny everything

搭配使用： filesystem

Run a skill in isolation to test it outside Claude Code

👤 Skill authors who want cross-agent validation ⏱ ~45 min intermediate

何时使用： You want to prove your skill works without the specific harness of Claude Code.

步骤

Load the skill

Load ~/.claude/skills/my-skill/ with openskills and print its metadata.✓ 已复制

→ Parsed frontmatter + body
Invoke with a prompt

Feed it a sample task and observe what would get loaded.✓ 已复制

→ Disclosure trace: what the agent would see

结果： Confidence your skill behaves portably.

注意事项

Skill depends on Claude Code internals — Refactor the skill to rely only on the public SKILL.md interface

组合

与其他 MCP 搭配，撬动十倍杠杆

openskills-skill + skill-optimizer-skill

Validate skills portably before optimizing them

Load the skill with openskills, then hand off to skill-optimizer for patch suggestions.✓ 已复制

openskills-skill + filesystem

Operate on the skills directory directly

List all SKILL.md in ~/.claude/skills/, summarize each.✓ 已复制

工具

此 MCP 暴露的能力

工具	输入参数	何时调用	成本
skill-loader	skills dir	Agent startup	minimal
progressive-disclosure	skill id + stage	During agent reasoning	minimal
sandbox-runner	script + policy	Any time a skill spawns a subprocess	script runtime

成本与限制

运行它的成本

API 配额: none
每次调用 Token 数: runtime is prompt-light; you control what's disclosed
费用: free, open source
提示: Keep progressive disclosure on — loading every script upfront blows context

安全

权限、密钥、影响范围

凭据存储： none at runtime level

数据出站： none beyond what scripts themselves do

WASM sandboxing is experimental — don't rely on it for adversarial scripts
Seatbelt profiles should be tightened per your use case

故障排查

常见错误与修复

Binding import fails

Ensure the native binary matches your platform; rebuild from source if needed.

Sandbox blocks expected syscalls

Iterate on the Seatbelt profile; log denials to find the right allow list.

替代方案

openskills 对比其他方案

替代方案	何时用它替代	权衡
Claude Code native skill loader	You're only running inside Claude Code	No cross-agent portability

openskills

为什么要用

核心特性

实时演示

实际使用效果

安装

选择你的客户端

使用场景

实战用法： openskills

How to embed Agent Skills into your own agent runtime

前置条件

步骤

注意事项

Run a skill in isolation to test it outside Claude Code

步骤

注意事项

组合

与其他 MCP 搭配，撬动十倍杠杆

工具

此 MCP 暴露的能力

成本与限制

运行它的成本

安全

权限、密钥、影响范围

故障排查

常见错误与修复

替代方案

openskills 对比其他方案

更多

资源