mcp-agent-security-gateway

poojakira/mcp-agent-security-gateway
★ 17 stars Python Code/Dev Tools Updated 3mo ago ⚠ Review
Monitor MCP tool calls for prompt injection, PII leakage, shadow servers, and exfiltration patterns.
View on GitHub → 🛡️ Security & TRiSM Audit →

⚡ What This Server Does & Key Capabilities

The mcp-agent-security-gateway Model Context Protocol (MCP) server provides seamless integration between AI assistants (such as Claude Desktop, Cursor, and Claude Code) and Monitor MCP tool calls for prompt injection, PII leakage, shadow servers, and exfiltration patterns. By connecting this server, your AI agent gains specialized capabilities to inspect, query, and automate tasks within the Code/Dev Tools domain.

Core Agent Capabilities

💡 Example Prompts to Ask Your AI Agent

Once installed in Claude Desktop or Cursor, try asking your assistant:

"Review recent commits in this repository, run the test suite, and highlight performance bottlenecks."
"Search for functions that don't have unit tests and generate test cases for them."

Compatible AI Clients

🟣 Claude Desktop ⚡ Cursor IDE 💻 Claude Code CLI 🌊 Windsurf 🤖 Gemini CLI

Quick Install

Copy the config for your editor. Verify environment variables or setup notes in the README.

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "mcp-agent-security-g": {
      "command": "uvx",
      "args": [
        "mcp-agent-security-gateway"
      ]
    }
  }
}

Or install with pip: pip install mcp-agent-security-gateway

⚡
Building Production MCP Agents? PRO TOOLKITS
Get the Complete 31,400+ MCP Ecosystem Dataset ($29) or the FastMCP + TypeScript Starter Kit ($39) with turnkey Docker & Cloudflare deployment.

README Documentation

Stdio proxy that intercepts MCP tools/call JSON-RPC messages and runs them through a multi-layer security pipeline. Returns Allow / Block / Redact verdicts. Sits between an MCP agent and tool server, inspects every tools/call request, and makes a security decision before anything reaches the downstream server. No code changes required on either side.

View Full Documentation & Updates on GitHub ↗

Topics

ai-securitydata-loss-preventionmcpprompt-injectionpython