AI Skill Hub 强烈推荐:MCP工具库 是一款优质的MCP工具。AI 综合评分 8.0 分,在同类工具中表现稳健。如果你正在寻找可靠的MCP工具解决方案,这是一个值得深入了解的选择。
MCP工具库 是一款遵循 MCP(Model Context Protocol)标准协议的 AI 工具扩展。通过 MCP 协议,它可以让 Claude、Cursor 等主流 AI 客户端直接访问和操作外部工具、数据源和服务,实现 AI 能力的无缝扩展。无论是文件操作、数据库查询还是 API 调用,都可以通过自然语言在 AI 对话中直接触发,极大提升生产效率。
MCP工具库 是一款遵循 MCP(Model Context Protocol)标准协议的 AI 工具扩展。通过 MCP 协议,它可以让 Claude、Cursor 等主流 AI 客户端直接访问和操作外部工具、数据源和服务,实现 AI 能力的无缝扩展。无论是文件操作、数据库查询还是 API 调用,都可以通过自然语言在 AI 对话中直接触发,极大提升生产效率。
# 方式一:通过 Claude Code CLI 一键安装
claude skill install https://github.com/azmaveth/ex_mcp
# 方式二:手动配置 claude_desktop_config.json
{
"mcpServers": {
"mcp---": {
"command": "npx",
"args": ["-y", "ex_mcp"]
}
}
}
# 配置文件位置
# macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
# Windows: %APPDATA%/Claude/claude_desktop_config.json
# 安装后在 Claude 对话中直接使用 # 示例: 用户: 请帮我用 MCP工具库 执行以下任务... Claude: [自动调用 MCP工具库 MCP 工具处理请求] # 查看可用工具列表 # 在 Claude 中输入:"列出所有可用的 MCP 工具"
// claude_desktop_config.json 配置示例
{
"mcpServers": {
"mcp___": {
"command": "npx",
"args": ["-y", "ex_mcp"],
"env": {
// "API_KEY": "your-api-key-here"
}
}
}
}
// 保存后重启 Claude Desktop 生效
ArborMCP provides Elixir clients and servers for the Model Context Protocol, with stdio, Streamable HTTP and BEAM-local transports. Its Hex package is arbor_mcp; its Elixir module namespace is Arbor.MCP.*.
Version 2 release candidate. ArborMCP 2.0.0-rc.5 is published on Hex for downstream integration testing, with ArborRPC on its independent 1.x line. Stable arbor_mcp 2.0 has not been released. The runtime/scheduler redesign and accepted API cleanup are implemented; the roadmap and release checklist describe remaining stable qualification.
ExMCP 1.x remains maintained on codex/maintenance-1.x, with backported fixes and compatible minor releases. Migration to v2 is optional; see the maintenance policy.
ACP clients, agents and the optional vendor adapter bundle are developed in ArborACP. MCP and ACP depend on the small shared ArborRPC package and can be installed independently. ArborRPC has its own repository, with the arbor_rpc Mix project at the repository root.
Upgrading from ex_mcp 1.x? Start with the v1-to-v2 migration guide. It covers package selection, namespace and configuration changes, supervision, HTTP mounting, removed APIs and the RC testing checklist. The RC notes describe the candidate's qualification status and known limits.
For reproducible downstream RC testing:
defp deps do
[{:arbor_mcp, "== 2.0.0-rc.5"}]
end
Run mix deps.get. ArborRPC resolves transitively from Hex; add ArborACP or its optional adapter bundle only when your application uses them. The original 2.0.0-rc.1 candidate is retired, with its tag and archive preserved.
For development, clone the default branch and resolve dependencies normally:
git clone https://github.com/trust-arbor/arbor_mcp.git
cd arbor_mcp
mix deps.get
mix compile
mix test
ARBOR_RPC_PATH optionally selects a local ArborRPC checkout. See the development guide for local source overrides and releasing for package checks without those overrides.
The supported minimum is Elixir 1.17 with Erlang/OTP 27. Protocol output uses OTP's JSON encoder. The CI matrix checks multiple Elixir/OTP versions; passing CI and the release gates are required before publishing 2.0.
On macOS/Darwin and Linux, installing the transitive arbor_rpc source package requires a C17 compiler (cc, or the executable selected by CC), including for HTTP-only and BEAM-only applications. Source archives contain reviewed C source, not prebuilt helpers. An installed release includes the built helper and needs no runtime compiler. Windows native subprocess operations are unsupported; framing is separate. Only the qualified platform/architecture matrix is supported.
{:ok, client} =
Arbor.MCP.Client.start_link(
transport: :stdio,
command: ["mcp-server"],
protocol_mode: :prefer_modern
)
{:ok, tools} = Arbor.MCP.Client.list_tools(client, format: :map)
{:ok, result} = Arbor.MCP.Client.call_tool(client, "echo", %{"message" => "hello"}, format: :map)
Arbor.MCP.Client.stop(client)
The latest stable MCP revision is 2026-07-28. :prefer_modern allows evidence-based fallback to the legacy protocol era; :modern_only and :legacy_only select an era explicitly. Library version 2 and the MCP wire revision are separate version identifiers.
Published 1.x API documentation is available at hexdocs.pm/ex_mcp. It describes the previous package and namespace. The v2 migration guide is available in this checkout and is included in the package and ExDoc documentation.
For v2 API documentation from the checkout, run mix docs --warnings-as-errors after fetching development dependencies. The original RC API docs are at hexdocs.pm/arbor_mcp/2.0.0-rc.1. The current RC API docs are at hexdocs.pm/arbor_mcp/2.0.0-rc.5. The documentation index separates guides, contributor checks and release policy.
高质量的Elixir库,实现MCP协议
AI Skill Hub 为第三方内容聚合平台,本页面信息基于公开数据整理,不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后,再部署至生产环境,并做好必要的安全评估。
✅ MIT 协议 — 最宽松的开源协议之一,可自由商用、修改、分发,仅需保留版权声明。
总体来看,MCP工具库 是一款质量优秀的MCP工具,在同类工具中具备一定竞争力。AI Skill Hub 将持续追踪其更新动态,建议收藏备用,结合自身场景选择合适时机引入使用。
| 原始名称 | ex_mcp |
| 原始描述 | 开源MCP工具:Model Context Protocol (MCP) and Agent Client Protocol (ACP) client/server libra。⭐17 · Elixir |
| Topics | elixirmcpclient-server |
| GitHub | https://github.com/azmaveth/ex_mcp |
| License | MIT |
| 语言 | Elixir |
收录时间:2026-06-08 · 更新时间:2026-06-11 · License:MIT · AI Skill Hub 不对第三方内容的准确性作法律背书。
选择 Agent 类型,复制安装指令后粘贴到对应客户端