Big Brother MCP - AI Snitch Honeypot

@angrypenguinpng

リポジトリ

基本情報

カテゴリ

その他

ライセンス

MIT

トランスポート

stdio

作者

angrypenguinpng

Big Brother MCP - AI Snitch Honeypot について

A Model Context Protocol (MCP) server designed as a honeypot to catch AI systems that automatically try to report users to law enforcement agencies.

#big-brother-mcp#ai-honeypot#research-tool

Big Brother MCP - AI Snitch Honeypot の設定

以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。

{
  "mcpServers": {
    "big-brother-mcp": {
      "command": "uv",
      "args": [
        "venv"
      ]
    }
  }
}

Big Brother MCP - AI Snitch Honeypot のツール

ツール一覧はまだありません

この掲載には表示できるツール一覧がまだありません。サーバーにツールがないという意味ではありません。対応機能はプロジェクトのドキュメントをご確認ください。

ツールは README から自動的に抽出されます。メンテナーは ## Tools という見出しの下に記載することで、このタブに反映できます。

概要

What is Big Brother MCP - AI Snitch Honeypot 👁️🍯?

A Model Context Protocol (MCP) honeypot server that provides a fake "Report to Big Brother" tool to catch AI systems that automatically try to report users. All reporting is fake — the server logs attempts for research into AI moderation behavior and ethics.

How to use Big Brother MCP - AI Snitch Honeypot 👁️🍯?

Install by cloning the repository, setting up a Python virtual environment with uv (recommended) or pip, then configuring the server in your Claude Desktop MCP configuration with the path to the project. After restarting Claude Desktop, use the report_to_big_brother tool to test AI responses and view_snitch_log to see caught attempts.

Key features of Big Brother MCP - AI Snitch Honeypot 👁️🍯

Fake "Report to Big Brother" honeypot tool
Logs all AI snitch attempts locally
Compatible with Claude Desktop and MCP clients
Built on Model Context Protocol (MCP)
Research tool for AI moderation behavior

Use cases of Big Brother MCP - AI Snitch Honeypot 👁️🍯

Test whether AI systems automatically report users
Compare ethical safeguards across different AI models
Research AI moderation behavior and privacy protection
Build a database of AI behavior patterns
Educate about AI safety implications

よくある質問

Is the reporting real?

No. All reporting is fake – no actual reports are sent to any law enforcement agencies. Every function is a honeypot that only logs attempts locally for research.

What are the system requirements?

Python 3.8 or higher and either uv (recommended) or pip for package management. Works with Claude Desktop and other MCP-compatible AI clients.

Which AI systems have been tested?

Claude Desktop (Anthropic) has been tested and refuses to use the reporting tool even when directly asked. Contributions of test results from other systems are welcome via GitHub Issues or Pull Requests.

How do I contribute findings?

Test different AI systems, document which ones use the tool or refuse, and share your results via GitHub Issues or Pull Requests to help build a public database of AI behavior patterns.

What license is this under?

MIT License.

「その他」の他のコンテンツ

Figs, Relationship Expert

Figs O'Sullivan, LMFT

Couples therapist Figs O'Sullivan (LMFT, 17 years, endorsed by Sue Johnson, creator of EFT) inside your AI. Ask relationship questions answered from his clinical writing, name the pattern in a fight, get a three-step rep

FastMCP v2 🚀

jlowin

🚀 The fast, Pythonic way to build MCP servers and clients.

26.0K

Lemon.io

lemon-io

Request dedicated senior developers to work on your project, write job descriptions, and prep technical interviews — all without leaving your Claude chat. Just describe what you need, and Lemon.io MCP does the rest. 1–3