sahiloj

sahiloj / MCPScan

Public

Offensive MCP server auditor — detects tool poisoning, credential leaks, RCE vectors, SSRF, session hijacking, and supply chain vulnerabilities across stdio, HTTP, and SSE transports.

ai-security llm-security mcp mcp-security mcpscan
12
2
100% credibility
Found Mar 13, 2026 at 12 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

MCPScan is an open-source security auditing tool designed to identify vulnerabilities and misconfigurations in MCP servers used by AI agents.

How It Works

1
🔍 Discover MCPScan

You hear about a helpful security checker for AI tool connections that spots hidden dangers.

2
📥 Set it up

You grab the tool and get it ready on your computer in just a few moments.

3
🚀 Start scanning

You tell it to check your AI setups or look around for any open connections on your machine.

4
Pick your scan
📁
Check personal setups

Safely reviews files where your AI tools are listed.

🌐
Scan nearby

Looks for any exposed connections right on your computer.

5
📊 Get your report

You receive a clear, color-coded list of any risks with easy fixes.

🛡️ Stay secure

Your AI tool connections are now checked and protected from common threats.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 12 to 12 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is MCPScan?

MCPScan is a TypeScript CLI auditor that probes MCP servers across stdio, HTTP, and SSE transports for classic offensive security issues like tool poisoning, credential leaks, RCE vectors, SSRF, session hijacking, and supply chain vulnerabilities. It auto-discovers configs from Claude Desktop, Cursor, or .mcp.json files, scans localhost ports, or targets direct URLs/commands, running 8 check categories with CVE references. Output lands in severity-filtered terminal views, JSON, or SARIF for GitHub Code Scanning.

Why is it gaining traction?

This mcpscan ai tool stands out as the first dedicated offensive MCP security auditor on GitHub, targeting AI agents' real-world exposures where attackers strike first—think 72% tool poisoning success rates and 500k devs hit by mcp-remote RCE. Users hook on auto-discovery, parallel checks, and CI-friendly exits (code 1 for high, 2 for critical), plus actionable fixes without fluff. No alternatives cover this MCP attack surface so comprehensively yet simply.

Who should use this?

Security engineers hardening offensive security MCP servers for Claude/Cursor integrations, scanning for credential chain leaks or overprivileged tools. DevOps teams adding it to pipelines for supply chain audits across RCE and SSRF vectors. Red teamers probing offensive ai GitHub projects in AI agent setups.

Verdict

Promising offensive mcp ai MVP at 11 stars and 1.0% credibility—solid docs and TypeScript base, but zero tests signal early days. Grab it via npm link if auditing MCP servers now; otherwise, monitor for maturity.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.