gaoharimran29-glitch/Cybersecurity-MCP-Server

What is this?

AI Clients by default has zero native cybersecurity tooling. No WHOIS. No DNS enumeration. No port scanning. No SSL inspection.

This MCP server fixes that — extending AI Clients with real-world security tools that run live against any domain or IP. Reconnaissance that normally requires multiple specialized tools and 20+ minutes of manual work becomes a single prompt.

This is a local MCP server — it runs entirely on your machine. Your data never leaves your computer. It is also listed on glama mcp registry.

---

Tools Available

Included in full_recon

Tool	Description
whois_lookup	Domain registration data — owner, registrar, creation date, expiry, name servers
dns_enumeration	A, AAAA, MX, NS, TXT, CNAME, SOA records + common subdomain brute-forcing
port_scan	Nmap-powered scanner with service/version detection and security warnings
ssl_inspect	SSL/TLS certificate — issuer, expiry, cipher strength, SANs, TLS version
email_security_check	Checks SPF, DKIM, and DMARC DNS records — returns a security_score, rating, and actionable recommendations for missing or weak configurations
tech_stack_detect	Web server, CMS, JS frameworks, CDN, analytics, and security header scoring
cert_transparency	Query crt.sh Certificate Transparency logs for a domain and extract certificate, subdomain, and wildcard information
asn_lookup	Autonomous System Number (ASN) and network ownership lookup via Team Cymru WHOIS — identifies ASN, BGP prefix, organization, registry, country, and allocation date for domains or IP addresses (no API key required)
ip_reputation	Check if an IP is flagged as malicious via AbuseIPDB (api key requied)
full_recon	Runs all core tools in parallel and returns combined result

Standalone Tools

Tool	Description
headers_analyzer	Analyzes HTTP security headers — checks HSTS, CSP, X-Frame-Options, and more with severity ratings and misconfiguration details
cve_lookup	Search NVD for known CVEs by software name and version (no API key required)
cloud_exposure_check	Checks for publicly accessible AWS S3, Azure Blob Storage, and Google Cloud Storage buckets using common bucket naming patterns derived from the target domain
trace_redirects	Traces the full HTTP redirect chain hop by hop — flags TLS downgrades, private-IP leaks, redirect loops, cross-domain hops, and overly long chains
robots_txt_inspect	Fetch and parse robots.txt to reveal hidden directories and sitemaps
hibp_check	Check if an email or domain appears in Have I Been Pwned breaches (HIBP_API_KEY required)

Prompts Available

Prompt	Description
threat_analysis	Generate a threat analysis prompt for the full_recon tool

---

📸 Demo

Single tool — CVE lookup

You: Look up CVEs for apache 2.4.49

Claude: Found 2 critical CVEs for Apache 2.4.49:
        CVE-2021-41773 (Score: 9.8 CRITICAL) — Path traversal vulnerability
        allowing remote code execution if CGI is enabled. Actively exploited
        in the wild...

Full recon

You: Do a complete security recon on reddit.com

Claude: [calls full_recon → runs all core tools in parallel → delivers full analysis]

---

📋 Prerequisites

• Python 3.12+ — download
• Claude Desktop — download or any other client you want
• Nmap — required for port scanning (download)
• Git — download
• uv - download

⚙️ Installation

Step 1 — Clone the repository

git clone https://github.com/AynOps/AynOps
cd AynOps

Step 2 — Create virtual environment and install dependencies

uv sync

If uv is not installed, install it using:

pip install uv

Step 3 — Install Nmap

Windows:

1. Download from nmap.org/download.html and run the installer
2. Manually add Nmap to PATH:
   • Press Win + S → search "Environment Variables"
   • Under System Variables → find Path → click Edit
   • Click New → add C:\Program Files (x86)\Nmap
   • Click OK on all windows
3. Restart your terminal and verify:

nmap --version

Mac:

brew install nmap

Linux:

sudo apt install nmap

Step 4 — Connect to Claude Desktop

Open your Claude Desktop config file:

OS	Path
Windows	%APPDATA%\Claude\claude_desktop_config.json
Mac	~/Library/Application Support/Claude/claude_desktop_config.json
Linux	~/.config/Claude/claude_desktop_config.json

Add this configuration:

Windows:

{
  "mcpServers": {
    "AynOps": {
      "command": "C:\\full\\path\\to\\AynOps\\.venv\\Scripts\\python.exe",
      "args": ["C:\\full\\path\\to\\AynOps\\server.py"],
      "env": {
        "ABUSEIPDB_API_KEY": "your-api-key-here",
        "HIBP_API_KEY": "your-hibp-api-key-here"
      }
    }
  }
}

Mac/Linux:

{
  "mcpServers": {
    "AynOps": {
      "command": "/full/path/to/AynOps/.venv/bin/python3",
      "args": ["/full/path/to/AynOps/server.py"],
      "env": {
        "ABUSEIPDB_API_KEY": "your-api-key-here",
        "HIBP_API_KEY": "your-hibp-api-key-here"
      }
    }
  }
}

⚠️ Always use the full absolute path to your .venv Python executable — not just python or python3. Claude Desktop may use a different Python installation otherwise.

Note: ABUSEIPDB_API_KEY is only required for the ip_reputation tool. Get a free key at abuseipdb.com. HIBP_API_KEY is only required for hibp_check; get a key at haveibeenpwned.com/API/Key.

Step 5 — Restart Claude Desktop

Fully quit and reopen Claude Desktop — closing the window is not enough. Check the system task manager and quit from there.

Verify tools are connected by asking Claude:

What cybersecurity tools do you have available?

Claude should list all tools.

---

Alternative — Run with Docker

The repository ships a Dockerfile that installs Nmap, uv, and project dependencies so you do not need a local Python virtualenv.

Build

git clone https://github.com/AynOps/AynOps
cd AynOps
docker build -t aynops .

Smoke-test the image (stdio MCP servers speak JSON-RPC on stdin/stdout; a one-shot help/version style check is enough to confirm the process starts):

docker run --rm -i aynops
# process waits for MCP JSON-RPC on stdin — Ctrl+C to exit

Claude Desktop config (Docker)

Claude Desktop launches MCP servers as local processes. Point command at docker and pass the image via args. Use absolute paths only if you mount files; for the stock image the container is self-contained.

OS	Path
Windows	%APPDATA%\Claude\claude_desktop_config.json
Mac	~/Library/Application Support/Claude/claude_desktop_config.json
Linux	~/.config/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "AynOps": {
      "command": "docker",
      "args": [
        "run",
        "--rm",
        "-i",
        "-e", "ABUSEIPDB_API_KEY=your-api-key-here",
        "-e" , "HIBP_API_KEY=your-api-key-here" ,
        "aynops"
      ]
    }
  }
}

Notes and caveats:

• stdio only — the container must stay attached (-i). Do not use -d.
• Nmap is inside the image — host Nmap is not required for the Docker path.
• Network scans need network access — default Docker bridge networking is enough for public internet targets. Host-network mode is optional and OS-specific.
• ABUSEIPDB_API_KEY is only required for ip_reputation. Omit the -e line if you do not use that tool.
• Rebuild after pulling updates: docker build -t aynops .
• On Windows, ensure Docker Desktop is running and docker is on PATH for the account that launches Claude Desktop.

---

📦 Listed On

Registry	Link
Official MCP Registry	registry.modelcontextprotocol.io
Glama	glama.ai/mcp/servers/...

🚀 Usage

Basic tool usage

Do a WHOIS lookup on example.com
Run DNS enumeration on github.com
Scan ports on scanme.nmap.org
Inspect the SSL certificate of stripe.com
Analyze HTTP security headers for github.com
Detect the tech stack of wordpress.org
Look up CVEs for apache 2.4.49
Look up CVEs for log4j 2.14.1
Check the reputation of IP 1.2.3.4
ASN Lookup for google.com
Does example.com have any exposed cloud storage?

Port scan types

Type	Description	Speed
basic	Top 100 ports	Fast (~5s)
service	Service & version detection	Medium (~15s)
os	OS detection (requires admin)	Medium
full	All 65535 ports	Slow (~5min)
vuln	Vulnerability scripts	Slow (~30s)

Scan scanme.nmap.org with service detection

Full recon

For getting best summarized result for the full_recon tool output, please use the prompt threat_analysis ( if your MCP client support ) by following these steps:

Note: Prompt support and automatic prompt chaining vary between MCP clients. Future client updates may improve this workflow.

For Claude Desktop:

1. Start Claude Desktop
2. Click on the + icon
3. Click on connectors
4. If the AynOps is connected, it will show the option for Add from AynOps.
5. Then choose the prompt threat_analysis
6. Claude will use the selected prompt to generate a correlated threat intelligence report.

After this ask this example prompt:

Do a complete security recon on scanme.nmap.org

Claude will run all core tools in parallel and deliver a full security analysis.

Follow-up analysis

Based on the recon, what are the top security risks?
What do the open ports mean from an attacker's perspective?
Is this SSL configuration strong enough for a financial services company?
Cross-reference the open ports with known CVEs for the detected services.

⚠️ Important: LLM Client Behavior

AynOps is an MCP server, which means tool execution is orchestrated by your MCP client (e.g., Claude Desktop, Cursor, VS Code, etc.), not by AynOps itself.

Tool execution may vary

Some LLM clients perform safety checks before invoking MCP tools. As a result, requests such as:

"Run a full security recon on example.com"

may be allowed in one conversation but declined in another, even when using the same server.

This behavior is determined by the LLM client, not by AynOps.

⚠️ Legal & Ethical Usage

Only scan domains and IPs you own or have explicit written permission to scan.

• All tools use public data — safe on any domain
• Port scanning should only target your own infrastructure or authorized systems
• The only public host officially permitted for Nmap testing is scanme.nmap.org
• Unauthorized port scanning may be illegal in your jurisdiction

Intended for:

• Security researchers
• Penetration testers (on authorized targets)
• Developers auditing their own infrastructure
• Students learning cybersecurity concepts

---

🗂️ Project Structure

├── .github/              # GitHub Actions workflows and templates
├── tests/                # Unit tests
├── tools/                # MCP tool implementations
  ├── prompts/            # AI prompt templates
  ├── signals/            # Signal extraction framework used by full_recon
    ├── registry.py       # Registers tools and their signal extractors
    └── extractor.py      # Executes all registered extractors and aggregates signals
├── utils/                # Shared helper utilities
├── server.py             # MCP server entry point
├── pyproject.toml        # Project metadata and dependencies
├── requirements.txt      # Python dependencies
├── mcp.json              # MCP server metadata
├── glama.json            # Glama MCP metadata
├── server.json           # MCP all tools metadata
├── Dockerfile            # Docker image definition
├── SECURITY.md           # Security policy
├── CONTRIBUTING.md       # Contribution guidelines
└── README.md             # Project documentation

---

Glama MCP Scoring

PyPI Package

mcp-name: io.github.AynOps/AynOps
Link:- https://pypi.org/project/AynOps/

🤝 Contributing

Pull requests are welcome! Check contributing.md for guidelines and a list of open issues ready to pick up.

---

📜 License

MIT License — free to use, modify, and distribute.

---

👤 Author

Built by Gaohar Imran

• GitHub: @gaoharimran29-glitch
• LinkedIn: Gaohar Imran

---

⭐ If this project helped you, consider giving it a star on GitHub!