automateyournetwork/nmap-service-detection

Service Detection & Vulnerability Scanning with nmap

How to Call the nmap MCP Tools

python3 $MCP_CALL "python3 -u $NMAP_MCP_SCRIPT" TOOL_NAME '{"param":"value"}'

When to Use

• Identify what software/version is running on an open port
• Fingerprint the OS of a network device or server
• Run targeted NSE scripts (SSL cert check, banner grab, protocol probe)
• Scan for known CVEs and common misconfigurations
• Full reconnaissance sweep of a single host or small range

Available Tools

| Tool | Purpose | Privileges | |------|---------|-----------| | nmap_service_detection | Service name + version on open ports (-sV) | none | | nmap_os_detection | OS fingerprinting (-O) | cap_net_raw | | nmap_script_scan | Run specific NSE scripts | none | | nmap_vuln_scan | Run the "vuln" NSE script category | none | | nmap_full_recon | SYN + service + OS + default scripts all-in-one | cap_net_raw |

Workflow: Service Identification

When asked "what's running on this host?" or "identify the services":

Step 1: Service Version Detection

python3 $MCP_CALL "python3 -u $NMAP_MCP_SCRIPT" nmap_service_detection '{"target":"192.168.1.1","ports":"common","intensity":7}'

Returns per-port: service name, product, version, CPE identifier.

Step 2: OS Fingerprinting

python3 $MCP_CALL "python3 -u $NMAP_MCP_SCRIPT" nmap_os_detection '{"target":"192.168.1.1"}'

Works best when the target has at least one open and one closed port.

Workflow: Security Assessment

When asked "check this host for vulnerabilities" or "security scan":

Step 1: Full Recon

Run the all-in-one audit sweep:

python3 $MCP_CALL "python3 -u $NMAP_MCP_SCRIPT" nmap_full_recon '{"target":"192.168.1.1","ports":"common"}'

This combines SYN scan + service detection + OS fingerprinting + default NSE scripts.

Step 2: Vulnerability Scan

Run the vuln NSE category for known CVEs:

python3 $MCP_CALL "python3 -u $NMAP_MCP_SCRIPT" nmap_vuln_scan '{"target":"192.168.1.1","ports":"common"}'

This is slow — use on specific targets, not wide ranges.

Step 3: Targeted Script Scans

Run specific NSE scripts for focused checks:

# SSL certificate inspection
python3 $MCP_CALL "python3 -u $NMAP_MCP_SCRIPT" nmap_script_scan '{"target":"192.168.1.1","scripts":"ssl-cert,ssl-enum-ciphers","ports":"443"}'

# HTTP title + headers
python3 $MCP_CALL "python3 -u $NMAP_MCP_SCRIPT" nmap_script_scan '{"target":"192.168.1.1","scripts":"http-title,http-headers","ports":"80,443,8080"}'

# Banner grabbing
python3 $MCP_CALL "python3 -u $NMAP_MCP_SCRIPT" nmap_script_scan '{"target":"192.168.1.1","scripts":"banner","ports":"1-1024"}'

# SMB enumeration
python3 $MCP_CALL "python3 -u $NMAP_MCP_SCRIPT" nmap_script_scan '{"target":"192.168.1.1","scripts":"smb-enum-shares,smb-os-discovery","ports":"445"}'

Tool Parameters

nmap_service_detection

• target (required): IP, hostname, or CIDR range
• ports (optional): Port range or "common" for top 1000 (default: "common")
• intensity (optional): Version detection aggressiveness 0-9 (default: 7)

nmap_os_detection

• target (required): Single IP or hostname (ranges don't work well)

nmap_script_scan

• target (required): IP, hostname, or CIDR range
• scripts (required): NSE script name(s), e.g. "ssl-cert", "http-title,http-headers", "banner"
• ports (optional): Port range or "common" (default: "common")

nmap_vuln_scan

• target (required): IP or hostname (keep scope tight)
• ports (optional): Port range or "common" (default: "common")

nmap_full_recon

• target (required): IP, hostname, or small CIDR range (/28 or smaller)
• ports (optional): Port range or "common" (default: "common")

Common NSE Script Names

| Script | Purpose | |--------|---------| | ssl-cert | Display SSL certificate details | | ssl-enum-ciphers | List supported SSL/TLS ciphers | | http-title | Grab HTML page title | | http-headers | Dump HTTP response headers | | http-methods | Check supported HTTP methods | | banner | Grab service banners | | smb-enum-shares | Enumerate SMB shares | | smb-os-discovery | Discover OS via SMB | | ssh-hostkey | Show SSH host keys | | dns-brute | DNS subdomain brute force | | ftp-anon | Check for anonymous FTP |

Output Format

All tools return structured JSON:

• scan_id — for retrieving results later
• per_host — per-host breakdown with open ports, services, versions
• os_detection — OS match name, accuracy, device type
• results / vulnerability_findings — script output organized by port

Important Rules

• OS detection requires at least one open and one closed port to fingerprint accurately
• Vuln scans are slow — target specific hosts, not wide ranges
• Full recon combines multiple scan types — takes longer but gives comprehensive results
• All scans respect the CIDR allowlist and are audit-logged
• Scan results persist and can be retrieved with nmap_list_scans / nmap_get_scan