mikuproject-mcp

mikuproject-mcp

MCP server adapter for mikuproject, providing tools, resources, and prompts to interact with mikuproject through the Model Context Protocol.

Category
访问服务器

README

mikuproject-mcp

mikuproject-mcp is a local MCP server adapter for mikuproject.

It exposes mikuproject operations to MCP clients as tools, resources, and prompts. The server is intended for local stdio use first. It does not call external AI services, does not start a network listener, and does not replace the upstream mikuproject product logic.

Status

This repository is currently a Node.js / TypeScript MCP server implementation. The Node package metadata is prepared for a future public package release, but the current validation path is still to build and run it from a local checkout.

Requirements

  • Node.js 20 or later
  • npm
  • Bundled or configured mikuproject runtime artifacts

The checked-in runtime artifacts under runtime/ are used by the MCP adapter for local execution.

Build

npm install
npm run build

The MCP server entrypoint after build is:

packages/node/dist/index.js

The package name prepared for the Node release is:

@igapyon/mikuproject-mcp-node

MCP Client Configuration

Use the built stdio entrypoint from your MCP client configuration.

Example:

{
  "mcpServers": {
    "mikuproject": {
      "command": "node",
      "args": ["packages/node/dist/index.js"]
    }
  }
}

If your MCP client runs from a different working directory, use paths that are valid for that client environment.

Release Tarball Usage

GitHub Releases may provide an npm package tarball asset named like:

igapyon-mikuproject-mcp-node-0.1.0.tgz

After downloading the tarball, install it globally:

npm install -g ./igapyon-mikuproject-mcp-node-0.1.0.tgz

Then configure your MCP client to run the installed command:

{
  "mcpServers": {
    "mikuproject": {
      "command": "mikuproject-mcp"
    }
  }
}

You can also run the release tarball directly with npm exec without a global install. Replace the version and URL with the release asset you want to use:

npm exec --yes --package=https://github.com/igapyon/mikuproject-mcp/releases/download/v0.1.0/igapyon-mikuproject-mcp-node-0.1.0.tgz -- mikuproject-mcp

Example MCP client configuration:

{
  "mcpServers": {
    "mikuproject": {
      "command": "npm",
      "args": [
        "exec",
        "--yes",
        "--package=https://github.com/igapyon/mikuproject-mcp/releases/download/v0.1.0/igapyon-mikuproject-mcp-node-0.1.0.tgz",
        "--",
        "mikuproject-mcp"
      ]
    }
  }
}

Runtime Configuration

By default, runtime artifacts are resolved from runtime/.

You can override them with environment variables:

MIKUPROJECT_MCP_RUNTIME_NODE=/path/to/mikuproject.mjs
MIKUPROJECT_MCP_WORKSPACE=/path/to/workspace

MIKUPROJECT_MCP_WORKSPACE controls where generated state, projections, exports, reports, summaries, and diagnostics are written. If it is not set, the server uses workplace/ under this repository.

Tools

The server exposes product-prefixed tools derived from the upstream CLI command tree.

Core state and AI workflow tools:

  • mikuproject.ai_spec
  • mikuproject.ai_detect_kind
  • mikuproject.state_from_draft
  • mikuproject.ai_export_project_overview
  • mikuproject.ai_export_task_edit
  • mikuproject.ai_export_phase_detail
  • mikuproject.ai_validate_patch
  • mikuproject.state_apply_patch
  • mikuproject.state_diff
  • mikuproject.state_summarize

Import, export, and report tools:

  • mikuproject.export_workbook_json
  • mikuproject.export_xml
  • mikuproject.export_xlsx
  • mikuproject.import_xlsx
  • mikuproject.report_wbs_markdown
  • mikuproject.report_mermaid

Resources

Common resource URIs include:

  • mikuproject://spec/ai-json
  • mikuproject://state/current
  • mikuproject://state/{name}
  • mikuproject://export/workbook-json
  • mikuproject://export/project-xml
  • mikuproject://export/project-xlsx
  • mikuproject://report/wbs-markdown
  • mikuproject://report/mermaid
  • mikuproject://summary/{operationId}
  • mikuproject://diagnostics/{operationId}

Tool results include generated artifact paths and, where applicable, product-specific resource URIs.

Prompts

The server provides small product-specific prompts:

  • mikuproject.create_project_draft
  • mikuproject.revise_state_with_patch
  • mikuproject.review_artifact_diagnostics

Prompt text refers to the AI specification resource instead of duplicating the full product specification.

Diagnostics and Outputs

Tool results are JSON text results with:

  • ok
  • operation
  • operationId
  • diagnostics
  • generated artifact references

Operation summaries and diagnostics are saved under the workspace and can be read through:

  • mikuproject://summary/{operationId}
  • mikuproject://diagnostics/{operationId}

Security Notes

This server is intended for trusted local use. It invokes local runtime artifacts and reads or writes local files based on tool arguments and workspace configuration.

Do not configure it with untrusted runtime artifacts or expose it as a hosted network service without a separate design for authentication, workspace isolation, upload handling, storage policy, cleanup, audit, and runtime isolation.

Developer Documentation

Developer-facing repository layout, implementation order, and contract notes are in docs/development.md and docs/miku-soft-50-mcp-design-v20260427.md.

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选