Token导航 LogoToken导航TokenDH.com
gibrun (Wirnat) logo
开发工具stdio官方级别未说明来源级核验

gibrun (Wirnat)

MCP Server

@modelcontextprotocol/inspector

gibRun MCP Server是一个为后端开发者设计的工具,提供端到端的API测试功能,集成PostgreSQL查询、HTTP请求测试和Go项目构建自动化,适用于开发调试和持续集成场景。

工具数

23

提示词数

0

GitHub Stars

1

资源数

0
API测试调试工具JavaScriptClaude后端开发Claude DesktopClaudeCursor

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

作者 / 组织

wirnat

提供方

wirnat

最后核验

2026/5/17 20:23

运行时

Node.js

快速接入

先看主来源和安装命令,再打开仓库或文档;下面只保留这个条目的关键接入事实。

命令预览

npx @modelcontextprotocol/inspector node build/index.js

详细介绍

gibRun MCP Server

MCP Server untuk membantu backend programmer dalam proses end-to-end API testing dengan integrasi PostgreSQL, HTTP requests, dan Go build automation.

Fitur

Tools yang Tersedia:

  1. postgres_query - Execute PostgreSQL queries

- Mendapatkan UID dari database - Verifikasi hasil API di database - Query data untuk testing

  1. http_request - Test API endpoints (seperti curl)

- Support semua HTTP methods (GET, POST, PUT, PATCH, DELETE) - Custom headers dan body - Tracking response time

  1. build_go_project - Build Go project

- Compile Go code dengan build flags - Output binary ke path tertentu - Error reporting yang detail

  1. run_go_command - Execute Go commands

- Run tests (go test) - Run aplikasi (go run) - Manage dependencies (go mod tidy)

  1. read_source_file - Baca source code

- Examine code sebelum fixing - Debug issues

  1. write_source_file - Write/update source code

- Fix code issues - Update implementation

  1. execute_shell_command - Execute arbitrary shell commands

- Custom operations - Cleanup tasks - Run custom scripts

  1. dap_restart - Restart VSCode debugger session 🔥 NEW

- Hot reload setelah fix code - Auto rebuild before restart - Seamless debugging workflow - Works dengan VSCode Go debugger

  1. dap_send_command - Send custom DAP commands

- Advanced debugger control - Set breakpoints programmatically - Evaluate expressions - Custom DAP operations

Debugger Tools (mirror external/mcp-go-debugger)

gibRun sekarang otomatis menjalankan proxy MCP untuk external/mcp-go-debugger, sehingga semua tool debugger Delve tersedia langsung di server ini:

  • launch – Jalankan binary Go dengan Delve dan mulai sesi debugging baru.
  • attach – Attach ke proses Go yang sudah berjalan berdasarkan PID.
  • debug – Compile & debug file Go tertentu (mirip dlv debug path/to/file.go).
  • debug_test – Build serta debug fungsi test tertentu dengan flag tambahan.
  • set_breakpoint – Pasang breakpoint pada file + nomor baris.
  • list_breakpoints – Lihat semua breakpoint yang aktif beserta statusnya.
  • remove_breakpoint – Hapus breakpoint berdasarkan ID Delve.
  • continue – Lanjutkan eksekusi sampai breakpoint berikutnya atau proses berakhir.
  • step – Step into baris atau fungsi berikutnya.
  • step_over – Step over ke baris berikut tanpa masuk ke fungsi.
  • step_out – Step keluar dari fungsi saat ini.
  • eval_variable – Evaluasi ekspresi/variable dengan kedalaman custom.
  • get_debugger_output – Ambil STDOUT/STDERR yang ditangkap oleh Delve beserta konteks eksekusi.
  • close – Tutup sesi debugging aktif dan hentikan server Delve internal.

Instalasi

Prerequisites

  • Node.js 18+
  • PostgreSQL (untuk database testing)
  • Go 1.20+ (untuk build automation dan agar proxy debugger bisa menjalankan Delve)
  • Delve (dlv) + dependencies dari external/mcp-go-debugger

Install Dependencies

npm install

Build

npm run build

Konfigurasi

Menambahkan ke Claude Desktop

Edit file konfigurasi Claude Desktop:

MacOS: ~/Library/Application Support/Claude/claude_desktop_config.json

Windows: %APPDATA%\Claude\claude_desktop_config.json

Tambahkan konfigurasi berikut:

{
  "mcpServers": {
    "gibrun": {
      "command": "node",
      "args": ["/path/to/gibRun/build/index.js"]
    }
  }
}

Atau menggunakan npx (setelah publish):

{
  "mcpServers": {
    "gibrun": {
      "command": "npx",
      "args": ["-y", "gibrun-mcp-server"]
    }
  }
}

Konfigurasi untuk Cursor

Jika menggunakan Cursor, tambahkan di .cursor/mcp_config.json:

{
  "mcpServers": {
    "gibrun": {
      "command": "node",
      "args": ["/Users/iturban/development/mcp/gibRun/build/index.js"]
    }
  }
}

Environment Variables untuk PostgreSQL

Anda dapat menyimpan kredensial database langsung di konfigurasi agen MCP sehingga tidak perlu mengetik connection_string setiap kali memanggil postgres_query. Server akan mencoba urutan berikut:

  1. Nilai argumen connection_string (jika tetap diberikan).
  2. Environment variable POSTGRES_CONNECTION_STRING.
  3. Kombinasi POSTGRES_USER, POSTGRES_PASSWORD, POSTGRES_HOST (default localhost), POSTGRES_PORT (default 5432), dan POSTGRES_DB.

Contoh konfigurasi Claude/Cursor:

"environment": {
  "POSTGRES_USER": "postgres",
  "POSTGRES_PASSWORD": "postgres",
  "POSTGRES_HOST": "localhost",
  "POSTGRES_PORT": "5432",
  "POSTGRES_DB": "hairkatz_0_0_1"
}

Setelah environment di-set, cukup kirim query:

Tool: postgres_query
Args:
  query: "SELECT * FROM users LIMIT 5"

Jika butuh database lain, override dengan connection_string langsung pada tool call.

Konfigurasi Go Debugger Proxy

Debugger tools di atas dijalankan oleh proses mcp-go-debugger yang dipanggil otomatis:

  1. gibRun mencoba menjalankan binary mcp-go-debugger yang tersedia di $PATH.
  2. Jika tidak ditemukan, server fallback ke go run ./cmd/mcp-go-debugger di folder external/mcp-go-debugger (pastikan Anda sudah menjalankan go mod download di sana minimal sekali).

Anda dapat mengoverride perilaku tersebut via environment variable pada konfigurasi MCP client:

"environment": {
  "GIBRUN_GO_DEBUGGER_COMMAND": "/abs/path/to/mcp-go-debugger",
  "GIBRUN_GO_DEBUGGER_ARGS": "--log --log-output=rpc",
  "GIBRUN_GO_DEBUGGER_CWD": "/Users/you/Development/mcp/gibrun/external/mcp-go-debugger"
}
  • GIBRUN_GO_DEBUGGER_COMMAND — Path ke executable alternatif (misalnya hasil go install).
  • GIBRUN_GO_DEBUGGER_ARGS — Argumen tambahan yang akan diparsing dengan pemisah spasi sederhana.
  • GIBRUN_GO_DEBUGGER_CWD — Working directory paksa untuk proses debugger.

Jika proxy gagal start (misalnya Go belum terinstall), gibRun akan tetap berjalan tetapi hanya menampilkan tools lokal.

Contoh Penggunaan

Workflow End-to-End Testing

Scenario: Testing user registration API

  1. Query database untuk get UID yang tersedia
Tool: postgres_query
Args:
  connection_string: "postgresql://user:pass@localhost:5432/mydb"
  query: "SELECT id FROM users WHERE email = $1"
  params: ["test@example.com"]
  1. Test API endpoint
Tool: http_request
Args:
  url: "http://localhost:8080/api/users"
  method: "POST"
  headers: {
    "Content-Type": "application/json",
    "Authorization": "Bearer token123"
  }
  body: {
    "email": "newuser@example.com",
    "name": "Test User"
  }
  1. Verify hasil di database
Tool: postgres_query
Args:
  connection_string: "postgresql://user:pass@localhost:5432/mydb"
  query: "SELECT * FROM users WHERE email = $1"
  params: ["newuser@example.com"]
  1. Jika ada error, read source code
Tool: read_source_file
Args:
  file_path: "/path/to/project/handlers/user.go"
  1. Fix code
Tool: write_source_file
Args:
  file_path: "/path/to/project/handlers/user.go"
  content: "package handlers\n\n// Fixed code here..."
  1. Rebuild project
Tool: build_go_project
Args:
  project_path: "/path/to/project"
  build_flags: "-v"
  1. Test lagi (ulangi step 2-3)

Testing dengan Multiple Assertions

AI dapat melakukan workflow otomatis:

  • Query database untuk prepare test data
  • Call API endpoint
  • Verify response status dan data
  • Check database untuk confirm changes
  • Jika gagal, analyze error, fix code, rebuild, dan test ulang
  • Loop sampai test pass

API Reference

postgres_query

Execute PostgreSQL query.

Parameters:

  • connection_string (required): PostgreSQL connection string
  • query (required): SQL query
  • params (optional): Query parameters array

Returns:

{
  "success": true,
  "rowCount": 1,
  "rows": [...],
  "fields": [...]
}

http_request

Make HTTP request.

Parameters:

  • url (required): Target URL
  • method (optional): HTTP method (GET, POST, PUT, PATCH, DELETE)
  • headers (optional): HTTP headers object
  • body (optional): Request body object
  • timeout (optional): Timeout in milliseconds

Returns:

{
  "success": true,
  "status": 200,
  "statusText": "OK",
  "headers": {...},
  "data": {...},
  "duration_ms": 123
}

build_go_project

Build Go project.

Parameters:

  • project_path (required): Path to Go project
  • build_flags (optional): Additional build flags
  • output_path (optional): Output binary path

Returns:

{
  "success": true,
  "stdout": "...",
  "stderr": "...",
  "message": "Build completed successfully"
}

run_go_command

Execute Go command.

Parameters:

  • project_path (required): Path to Go project
  • command (required): Go command (e.g., "test ./...", "run main.go")

Returns:

{
  "success": true,
  "stdout": "...",
  "stderr": "...",
  "command": "go test ./..."
}

read_source_file

Read source code file.

Parameters:

  • file_path (required): Path to file

Returns:

{
  "success": true,
  "file_path": "...",
  "content": "...",
  "size": 1234
}

write_source_file

Write source code file.

Parameters:

  • file_path (required): Path to file
  • content (required): File content

Returns:

{
  "success": true,
  "file_path": "...",
  "size": 1234,
  "message": "File written successfully"
}

execute_shell_command

Execute shell command.

Parameters:

  • command (required): Shell command
  • working_dir (optional): Working directory

Returns:

{
  "success": true,
  "command": "...",
  "stdout": "...",
  "stderr": "..."
}

dap_restart

Restart VSCode debugger session via Debug Adapter Protocol.

Parameters:

  • port (optional): DAP server port (lihat di VSCode debug console: "DAP server listening at: 127.0.0.1:PORT"). Jika dikosongkan, gibRun akan mencari port otomatis memakai lsof -i -P -n | grep "dlv.*LISTEN" dan memverifikasi proses dlv dap.
  • host (optional): DAP server host (default: "127.0.0.1"). Ikut terdeteksi otomatis jika port dikosongkan.
  • rebuild_first (optional): Rebuild project before restart (default: true)
  • project_path (required if rebuild_first=true): Path to Go project

Returns:

{
  "success": true,
  "message": "Debugger restarted successfully",
  "dap_response": {...},
  "build_result": {...},
  "dap_server": "127.0.0.1:49279"
}

Example Usage:

AI Prompt: "Fix the bug di user_handler.go line 45, 
           rebuild, dan restart debugger di port 49279"

AI akan:
1. Read source file
2. Identify bug
3. Write fixed code
4. Use dap_restart dengan port 49279 
   (automatically rebuilds dan restarts debugger)

Auto-detecting DAP Address:

  • Tidak perlu mengisi port jika hanya ada satu proses dlv dap yang LISTEN — gibRun akan menjalankan perintah berikut dan memakai host/port yang ditemukan:
  lsof -i -P -n | grep "dlv.*LISTEN" | while read line; do 
    pid=$(echo "$line" | awk '{print $2}')
    if ps -p $pid -o command= 2>/dev/null | grep -q "dlv dap"; then
      echo "$line" | awk '{print "Port:", $9, "PID:", $2}'
    fi
  done
  • Jika lebih dari satu proses ditemukan, gibRun akan menampilkan daftar port tersebut dan meminta kamu memilih dengan mengisi port (dan opsional host).
  • Manual fallback: jalankan debugger VSCode (F5), buka Debug Console, lalu cari pesan DAP server listening at: 127.0.0.1:XXXXX dan masukkan port tersebut.

dap_send_command

Send custom DAP command for advanced debugger control.

Parameters:

  • port (optional): DAP server port. Kosongkan untuk auto-detect seperti di atas.
  • command (required): DAP command name (e.g., "restart", "disconnect", "evaluate")
  • host (optional): DAP server host (default: "127.0.0.1"). Ikut terdeteksi otomatis jika port tidak diisi.
  • arguments (optional): Command arguments as object

Returns:

{
  "success": true,
  "command": "evaluate",
  "response": {...},
  "dap_server": "127.0.0.1:49279"
}

Example Commands:

  • restart - Restart debugging session
  • disconnect - Stop debugging
  • evaluate - Evaluate expression
  • setBreakpoints - Set breakpoints programmatically

Development

Watch mode untuk development

npm run dev

Testing MCP Server

Gunakan MCP Inspector untuk testing:

npx @modelcontextprotocol/inspector node build/index.js

Keuntungan Menggunakan gibRun

  1. Automated Testing Workflow: AI dapat menjalankan complete test cycle secara otomatis
  2. Database Integration: Direct access ke PostgreSQL untuk verify data
  3. Quick Iteration: Auto-rebuild dan re-test ketika menemukan errors
  4. Seamless Development: Integrated dengan VSCode debugger workflow
  5. Smart Error Handling: AI dapat analyze errors dan propose fixes

Troubleshooting

Database Connection Issues

Pastikan PostgreSQL running dan connection string benar:

postgresql://username:password@host:port/database

Build Failures

Cek Go installation:

go version

Pastikan semua dependencies tersedia:

go mod download

Permission Issues

Pastikan file permissions correct untuk read/write operations.

License

MIT

Contributing

Contributions welcome! Please submit issues atau pull requests.

目录标签

目录标签

API测试调试工具JavaScriptClaude后端开发本地部署PostgreSQL集成Go构建自动化

支持客户端

Claude DesktopClaudeCursor

接入字段

传输方式(transport,传输协议)

stdio

鉴权方式(authType,认证方式)

token

运行时(runtime,运行环境)

Node.js

来源包(packageName,安装包名)

@modelcontextprotocol/inspector

工具数量(toolCount,工具数)

23

资源数量(resourceCount,资源数)

0

提示词数量(promptCount,提示词数)

0

权限和风险

stdiotoken部署方式未说明

接入前请确认传输方式、认证方式和部署位置,并根据实际工具能力限制访问范围。

安装前确认

不要直接授予不必要的文件、网络或账号权限;先核对安装命令和配置内容。

来源信息

继续浏览同类 MCP