Observability, security audit, and governance layer for Model Context Protocol (MCP) servers. Validates registrations, scans tool surfaces for prompt injection and PII, detects schema drift, and produces posture scores with policy outcome matrices. Enterprise platform-engineering doctrine, not a research demo.

Quick Install

Copy the config for your editor. Some servers may need additional setup — check the README.

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "mcp-sentinel": {
      "command": "npx",
      "args": [
        "-y",
        "mizcausevic-dev/mcp-sentinel"
      ]
    }
  }
}

README Excerpt

> 🛡️ **Live preview:** [mcp.kineticgain.com](https://mcp.kineticgain.com) — interactive governance dashboard for this engine. See [mcp-sentinel-dashboard](https://github.com/mizcausevic-dev/mcp-sentinel-dashboard) for the UX repo. TypeScript observability, security audit, and governance portfolio project demonstrating MCP server registration validation, schema-drift detection, prompt-injection scanning, PII guardrails, and posture scoring for enterprise AI platforms running Model Context Protoco

Topics

ai-governanceai-securityexpressmcpmodel-context-protocolplatform-engineeringprompt-injectionswaggertypescriptzod

mcp-sentinel

Quick Install

README Excerpt

Topics

Related Security Servers