Skip to main content
Glama
frida-mcp
by SnowLuma

frida-mcp

極めてシンプルなFrida MCPサーバー。

設計理念

  • Attach → LoadScript → Call/Listen → Unload の最短ワークフロー

  • Fridaのコーパスは多くのモデルで広く学習済み そのまま利用可能

  • ツール数を最小化 6つのツールのみを公開。少ないことは良いことであり、コンテキストを簡潔に保つ

Related MCP server: Frida MCP

ツール一覧

ツール

用途

attach

Fridaをプロセス(PIDまたはプロセス名)にアタッチする。プロセスがデタッチ/クラッシュした場合は再呼び出しが必要

detach

現在のプロセスから能動的に切断し、すべての永続スクリプトをアンロードする。ターゲットを変更する前に呼び出すのに適している

load_script

永続的なFrida JSを読み込む。rpc.exports = {...} を定義でき、send(data) でログを出力可能

call_rpc

load_script で登録されたRPCメソッドを呼び出し、結果を同期的に返す

get_script_output

スクリプトの send(...) メッセージキューを取得してクリアする(tail logのようなもの)

unload_script

永続スクリプトをアンロードしてクリーンアップする

インストール

ソースから:

npm install
npm run build

またはReleaseからダウンロード: GitHub Releases でプラットフォームに対応する圧縮ファイル(win-x64 / linux-x64 / macos-arm64)を選択し、解凍して使用します。npm install は不要です:

# 解压后会得到 frida-mcp/ 目录
node frida-mcp/dist/index.js

使用方法

直接起動(オプションでPIDを事前アタッチ可能)

# 不带参数启动,由 LLM 通过 attach 工具连接
node dist/index.js

# 或启动时直接 attach 到 PID
node dist/index.js 12345

MCPクライアント設定

{
  "mcpServers": {
    "frida": {
      "command": "node",
      "args": ["e:/mcp-ida-frida/dist/index.js"]
    }
  }
}

Thanks

  • Claude - AIプログラミングアシスタント

  • Frida - 動的インストゥルメンテーションフレームワーク

  • MCP - Model Context Protocol

Available Tools

6 tools
attachAttach to ProcessA

Attach Frida to a process. AI MUST call this first, and MUST call it again whenever any other tool reports that the process is detached or crashed.

Inputs: target (required).

call_rpcCall RPC MethodA

Call a function exported by a persistent script loaded via load_script. If this fails because the process is detached/crashed, call the attach tool again and then load_script before retrying.

Inputs: script (required), method (required), args.

detachDetach from ProcessA

Detach Frida from the current process and unload all persistent scripts. Call this when finished with the target or before attaching to a different process. After detach you MUST call attach again before any other tool.

Inputs: no named parameters.

get_script_outputGet Script OutputA

Retrieve accumulated messages (send(data)) from a persistently loaded script and clear its message queue. Similar to tailing a log.

Inputs: name (required).

load_scriptLoad Persistent ScriptA

Load a persistent background Frida script. Use this to set up hooks that collect data over time or long-running RPC exports. Afterwards, use call_rpc to invoke functions or get_script_output to check logs. If this fails because the process is detached/crashed, call the attach tool again and retry.

Inputs: name (required), source (required).

unload_scriptUnload Persistent ScriptC

Securely unload and clean up a persistent script.

Inputs: name (required).

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A4/5.0

    Scored across 6 tools

    Disambiguation5/5

    Each tool has a clear, distinct purpose: attach/detach for lifecycle, load_script/unload_script for script management, call_rpc for function invocation, and get_script_output for log retrieval. There is no overlap or ambiguity between tools.

    Naming Consistency5/5

    All tool names follow a consistent verb_noun pattern in snake_case (e.g., attach, load_script, call_rpc), making them predictable and easy to understand.

    Tool Count5/5

    With 6 tools, the set is well-scoped for the purpose of Frida process attachment and persistent script management. Each tool serves a necessary step in the workflow without being excessive or insufficient.

    Completeness4/5

    The tool set covers the core workflow of attaching, detaching, loading/unloading scripts, calling RPC, and retrieving output. However, it lacks support for ad-hoc scripts or advanced Frida operations, which might be a minor gap for some use cases.

    Maintenance

    ActivityNo data
    ResponsivenessOutdated data

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      D
      maintenance
      An MCP-compliant server that enables AI systems to interact with mobile and desktop applications through Frida's dynamic instrumentation capabilities, allowing for process management, device control, JavaScript execution, and script injection.
      468
      MIT
    • A
      license
      B
      quality
      Not graded
      maintenance
      Enables AI systems to interact with mobile and desktop applications through Frida's dynamic instrumentation capabilities. Provides process management, device control, JavaScript REPL execution, and script injection for runtime analysis and reverse engineering.
      14
      2
      MIT
    • A
      license
      Not graded
      quality
      D
      maintenance
      Enables dynamic instrumentation of running processes through Frida, allowing users to hook functions, inspect memory, modify process behavior, and execute JavaScript code for debugging and reverse engineering.
      3
      MIT
    • A
      license
      C
      quality
      D
      maintenance
      A Frida MCP server for authorized dynamic analysis and application security research, exposing device/session management, script injection, process and memory inspection, and platform-specific workflows.
      2
      100
      4
      MIT