
Mcp Security Hub
- 960 installs
- 10 repo stars
- Updated August 4, 2026
- aradotso/security-skills
mcp-security-hub is a Claude Code skill that deploys and orchestrates 38 Dockerized MCP servers exposing offensive security tools like Nmap, Nuclei, Ghidra, and SQLMap for developers who want AI agents to run security as
About
mcp-security-hub is a production-ready security skill from aradotso/security-skills bundling 38 Dockerized MCP servers for offensive security tooling. It exposes scanners and analysis tools—including Nmap, Nuclei, Ghidra, SQLMap, Radare2, and Gitleaks—so Claude and compatible agents invoke them via Model Context Protocol. Developers reach for mcp-security-hub when setting up AI-driven web security assessments, network scans, binary analysis, or secret scanning without manually chaining CLI commands. The skill covers deployment, orchestration, and integration of security tools as MCP servers for authorized testing workflows.
- 38 Dockerized MCP servers exposing 300+ offensive security tools
- 8 reconnaissance servers including Nmap, Shodan, Masscan and ProjectDiscovery tools
- 6 web security servers covering Nuclei, SQLMap, Nikto, ffuf and Burp Suite
- 6 binary analysis servers for radare2, Ghidra, Binwalk, YARA, Capa and IDA Pro
- Additional coverage for blockchain, cloud, secrets detection, fuzzing, OSINT and password cracking
Mcp Security Hub by the numbers
- 960 all-time installs (skills.sh)
- +110 installs in the week ending Aug 4, 2026 (Skillselion tracking)
- Ranked #405 of 2,203 Security skills by installs in the Skillselion catalog
- Data as of Aug 5, 2026 (Skillselion catalog sync)
npx skills add https://github.com/aradotso/security-skills --skill mcp-security-hubAdd your badge
Show developers this skill is listed on Skillselion. Paste this into your README.
| Installs | 960 |
|---|---|
| repo stars | ★ 10 |
| Last updated | August 4, 2026 |
| Repository | aradotso/security-skills ↗ |
How do you expose security tools as MCP servers?
Expose 38 offensive security tools as MCP servers that Claude can call directly in natural language.
Who is it for?
Security engineers and developers integrating authorized offensive tooling into Claude or MCP agent workflows via Docker.
Skip if: Teams needing compliance-only policy checklists without hands-on scanning or unauthorized penetration testing targets.
When should I use this skill?
User wants to set up, deploy, or run offensive security tools like Nmap, Nuclei, or Gitleaks through MCP servers in Claude.
What you get
Running Dockerized MCP security servers for network scans, vuln scanning, binary analysis, and secret detection
- MCP security server deployment
- scan and assessment results
By the numbers
- Bundles 38 Dockerized MCP servers for offensive security tools
Files
mcp-security-hub
Skill by ara.so — Security Skills collection.
Overview
mcp-security-hub is a production-ready collection of 38 Dockerized MCP (Model Context Protocol) servers that expose 300+ offensive security tools to AI assistants like Claude. It enables natural language security assessments, vulnerability scanning, binary analysis, and penetration testing workflows.
Key capabilities:
- 8 reconnaissance servers (Nmap, Shodan, ProjectDiscovery tools, WhatWeb, Masscan, ZoomEye)
- 6 web security servers (Nuclei, SQLMap, Nikto, ffuf, Burp Suite)
- 6 binary analysis servers (radare2, Ghidra, Binwalk, YARA, Capa, IDA Pro)
- 3 blockchain security servers (DAML Viewer, Medusa, Solazy)
- 3 cloud security servers (Trivy, Prowler, RoadRecon)
- Plus: secrets detection, fuzzing, OSINT, threat intelligence, Active Directory, password cracking
Installation
Prerequisites
- Docker 20.10+
- Docker Compose 2.0+
- Claude Desktop or MCP-compatible client
Clone and Build
git clone https://github.com/FuzzingLabs/mcp-security-hub.git
cd mcp-security-hub
# Build all MCP servers
docker-compose build
# Or build specific servers
docker-compose build nmap-mcp nuclei-mcp gitleaks-mcpVerify Installation
# Check built images
docker images | grep mcp
# Start specific servers
docker-compose up nmap-mcp nuclei-mcp -d
# Verify health
docker-compose psConfiguration
Claude Desktop Integration
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"nmap": {
"command": "docker",
"args": ["run", "-i", "--rm", "--cap-add=NET_RAW", "nmap-mcp:latest"]
},
"nuclei": {
"command": "docker",
"args": ["run", "-i", "--rm", "nuclei-mcp:latest"]
},
"gitleaks": {
"command": "docker",
"args": [
"run", "-i", "--rm",
"-v", "${HOME}/repos:/app/target:ro",
"gitleaks-mcp:latest"
]
},
"radare2": {
"command": "docker",
"args": [
"run", "-i", "--rm",
"-v", "${HOME}/binaries:/samples:ro",
"radare2-mcp:latest"
]
},
"sqlmap": {
"command": "docker",
"args": ["run", "-i", "--rm", "sqlmap-mcp:latest"]
},
"trivy": {
"command": "docker",
"args": [
"run", "-i", "--rm",
"-v", "/var/run/docker.sock:/var/run/docker.sock:ro",
"trivy-mcp:latest"
]
}
}
}Project-Level Configuration
Create .mcp.json in your project root:
{
"mcpServers": {
"nmap": {
"command": "docker",
"args": ["run", "-i", "--rm", "--cap-add=NET_RAW", "nmap-mcp:latest"]
},
"nuclei": {
"command": "docker",
"args": ["run", "-i", "--rm", "nuclei-mcp:latest"]
}
}
}Environment Variables
Many MCP servers require API keys for external services:
# Shodan
export SHODAN_API_KEY=your_key_here
# VirusTotal
export VT_API_KEY=your_key_here
# ZoomEye
export ZOOMEYE_API_KEY=your_key_here
# Burp Suite
export BURP_API_KEY=your_key_herePass environment variables to Docker containers:
{
"mcpServers": {
"shodan": {
"command": "docker",
"args": [
"run", "-i", "--rm",
"-e", "SHODAN_API_KEY=${SHODAN_API_KEY}",
"shodan-mcp:latest"
]
}
}
}Key MCP Servers
Nmap MCP (Network Scanning)
Available tools (8):
scan_hosts- Basic host discoveryscan_ports- Port scanning with service detectionscan_os- OS fingerprintingscan_vuln- Vulnerability scanning with NSE scriptsscan_custom- Custom nmap command executionlist_nse_scripts- List available NSE scriptsget_nse_script_info- Get NSE script detailsscan_with_script- Run specific NSE script
Example prompts:
- "Scan 192.168.1.0/24 for open ports"
- "Perform OS detection on 10.0.0.1"
- "Run vulnerability scan on example.com"
Nuclei MCP (Vulnerability Scanning)
Available tools (7):
scan_target- Scan with default templatesscan_with_severity- Filter by severity (critical, high, medium, low)scan_with_tags- Use specific tags (cve, exposure, xss, sqli)scan_with_templates- Use custom template pathslist_templates- Show available templatesupdate_templates- Update template databasescan_multiple_targets- Bulk scanning
Example prompts:
- "Scan https://example.com for critical vulnerabilities"
- "Check example.com for CVEs using nuclei"
- "Run nuclei with exposure and misconfiguration templates"
Gitleaks MCP (Secrets Detection)
Available tools (5):
scan_repo- Scan git repositoryscan_file- Scan individual filescan_directory- Scan directory treegenerate_baseline- Create baseline for false positivesscan_commits- Scan specific commit range
Example prompts:
- "Scan /app/target/myrepo for secrets"
- "Check this project for exposed API keys"
- "Find credentials in the last 10 commits"
Volume mounting required:
{
"gitleaks": {
"command": "docker",
"args": [
"run", "-i", "--rm",
"-v", "/path/to/repos:/app/target:ro",
"gitleaks-mcp:latest"
]
}
}Radare2 MCP (Binary Analysis)
Available tools (32+):
analyze_binary- Load and analyze binarydisassemble- Disassemble functionsdecompile- Decompile to C-like codelist_functions- Show all functionsfind_strings- Extract stringsfind_imports- List imported functionsfind_exports- List exported functionssearch_bytes- Search byte patternsanalyze_entropy- Detect packed sections
Example prompts:
- "Analyze /samples/malware.exe for suspicious functions"
- "Decompile main function in this binary"
- "Find strings in /samples/firmware.bin"
Volume mounting required:
{
"radare2": {
"command": "docker",
"args": [
"run", "-i", "--rm",
"-v", "/path/to/binaries:/samples:ro",
"radare2-mcp:latest"
]
}
}SQLMap MCP (SQL Injection)
Available tools (8):
test_url- Test URL for SQL injectiondump_database- Extract database contentsdump_table- Extract specific tablelist_databases- Enumerate databaseslist_tables- Enumerate tablesget_dbs- Get database namesget_current_user- Get DB usertest_forms- Test web forms for SQLi
Example prompts:
- "Test https://example.com/page?id=1 for SQL injection"
- "Dump database from vulnerable URL"
- "Check this form for SQL injection vulnerabilities"
Trivy MCP (Container Security)
Available tools (7):
scan_image- Scan Docker imagescan_filesystem- Scan local filesystemscan_config- Scan IaC files (Terraform, K8s)scan_repo- Scan git repositorylist_vulnerabilities- Show known CVEsget_sbom- Generate SBOMscan_kubernetes- Scan K8s cluster
Example prompts:
- "Scan nginx:latest for vulnerabilities"
- "Check this Dockerfile for security issues"
- "Generate SBOM for python:3.11 image"
Common Workflows
Network Reconnaissance Workflow
# 1. Build reconnaissance servers
docker-compose build nmap-mcp whatweb-mcp masscan-mcp
# 2. Start services
docker-compose up nmap-mcp whatweb-mcp -d
# 3. Use in Claude
# "Scan 10.0.0.0/24 for web servers, then fingerprint each one"Web Application Security Assessment
# Build web security stack
docker-compose build nuclei-mcp sqlmap-mcp ffuf-mcp
# Start services
docker-compose up nuclei-mcp sqlmap-mcp ffuf-mcp -d
# In Claude:
# "Scan example.com with nuclei, test any forms for SQL injection,
# and fuzz for hidden directories"Binary Analysis Pipeline
# Build binary analysis tools
docker-compose build radare2-mcp binwalk-mcp yara-mcp capa-mcp
# Mount binaries directory
docker-compose up radare2-mcp binwalk-mcp yara-mcp capa-mcp -d
# In Claude:
# "Analyze /samples/suspicious.exe - extract filesystem if packed,
# scan for malware patterns, and identify capabilities"Secrets Scanning in CI/CD
# Build gitleaks
docker-compose build gitleaks-mcp
# Run as one-off scan
docker run -i --rm \
-v "$(pwd):/app/target:ro" \
gitleaks-mcp:latest <<EOF
{
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"name": "scan_directory",
"arguments": {
"path": "/app/target"
}
},
"id": 1
}
EOFCloud Security Audit
# Build cloud security tools
docker-compose build trivy-mcp prowler-mcp
# Mount Docker socket for Trivy
docker-compose up trivy-mcp prowler-mcp -d
# In Claude:
# "Scan all running containers for CVEs, then audit AWS account
# for security misconfigurations"Docker Compose Orchestration
Start All Services
docker-compose up -dStart Specific Category
# Reconnaissance only
docker-compose up nmap-mcp whatweb-mcp masscan-mcp -d
# Web security only
docker-compose up nuclei-mcp sqlmap-mcp ffuf-mcp -dResource Limits
Edit docker-compose.yml to adjust resource constraints:
services:
nmap-mcp:
image: nmap-mcp:latest
deploy:
resources:
limits:
cpus: '2.0'
memory: 1G
reservations:
cpus: '0.5'
memory: 256MHealth Monitoring
# Check health status
docker-compose ps
# View logs
docker-compose logs -f nmap-mcp
# Restart unhealthy services
docker-compose restart nmap-mcpDevelopment
Building Individual Servers
cd reconnaissance/nmap-mcp
docker build -t nmap-mcp:latest .Testing MCP Server
# Run interactive test
docker run -it --rm nmap-mcp:latest
# Send JSON-RPC request
echo '{"jsonrpc":"2.0","method":"tools/list","id":1}' | \
docker run -i --rm nmap-mcp:latestAdding Custom MCP Server
mkdir -p custom-category/mytool-mcp
cd custom-category/mytool-mcp
# Create Dockerfile
cat > Dockerfile <<'EOF'
FROM python:3.11-slim
RUN useradd -m -u 1000 mcpuser
RUN pip install mcp mytool
USER mcpuser
WORKDIR /app
COPY server.py .
CMD ["python", "server.py"]
EOF
# Create server.py with MCP protocol implementation
# Add to docker-compose.ymlSecurity Hardening
All MCP servers follow security best practices:
# Example hardened Dockerfile pattern
FROM alpine:3.19
RUN adduser -D -u 1000 mcpuser
RUN apk add --no-cache tool-name
USER mcpuser
WORKDIR /app
# Drop all capabilities by default
# Add only required capabilities in docker-compose.ymlRequired Capabilities
Some tools need specific Linux capabilities:
nmap-mcp:
cap_drop:
- ALL
cap_add:
- NET_RAW # Required for SYN scanning
trivy-mcp:
volumes:
- /var/run/docker.sock:/var/run/docker.sock:ro # Docker scanningRead-Only Mounts
Always mount target directories read-only:
gitleaks-mcp:
volumes:
- ./repos:/app/target:ro # Read-only prevents modificationTroubleshooting
MCP Server Not Responding
# Check if container is running
docker ps | grep mcp
# View logs
docker logs nmap-mcp
# Restart service
docker-compose restart nmap-mcp
# Test JSON-RPC directly
echo '{"jsonrpc":"2.0","method":"tools/list","id":1}' | \
docker run -i --rm nmap-mcp:latestPermission Denied Errors
# Nmap requires NET_RAW capability
# Add to docker-compose.yml:
cap_add:
- NET_RAW
# Or run with --cap-add
docker run --cap-add=NET_RAW nmap-mcp:latestVolume Mount Issues
# Ensure absolute paths
docker run -v /absolute/path:/app/target:ro gitleaks-mcp
# Check permissions (container runs as UID 1000)
chown -R 1000:1000 /path/to/repos
# Verify mount inside container
docker run -it --rm -v $(pwd):/app/target:ro gitleaks-mcp sh
ls -la /app/targetClaude Desktop Not Finding MCP Servers
# Verify config location
# macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
# Windows: %APPDATA%\Claude\claude_desktop_config.json
# Check JSON syntax
cat ~/Library/Application\ Support/Claude/claude_desktop_config.json | jq .
# Restart Claude Desktop after config changes
# Verify image exists
docker images | grep nmap-mcpAPI Key Authentication Failures
# Verify environment variable is set
echo $SHODAN_API_KEY
# Pass to Docker container
docker run -e SHODAN_API_KEY=$SHODAN_API_KEY shodan-mcp
# For Claude Desktop, use full env var syntax
{
"command": "docker",
"args": ["-e", "SHODAN_API_KEY=${SHODAN_API_KEY}", ...]
}Network Connectivity Issues
# Some tools need host network access
docker run --network host nmap-mcp
# Or create custom network
docker network create security-net
docker run --network security-net nmap-mcpContainer Build Failures
# Clear Docker build cache
docker builder prune -a
# Rebuild with no cache
docker-compose build --no-cache nmap-mcp
# Check base image availability
docker pull alpine:3.19
docker pull python:3.11-slimAdvanced Usage
Custom Nuclei Templates
# Mount custom template directory
docker run -i --rm \
-v $(pwd)/custom-templates:/nuclei-templates:ro \
nuclei-mcp:latest
# In Claude: "Use custom nuclei templates from /nuclei-templates"Multi-Stage Binary Analysis
# 1. Extract firmware
docker run -v $(pwd)/firmware:/samples:ro binwalk-mcp
# 2. Scan extracted files
docker run -v $(pwd)/firmware/_extracted:/samples:ro yara-mcp
# 3. Analyze suspicious binaries
docker run -v $(pwd)/firmware/_extracted:/samples:ro radare2-mcpAutomated Scanning Pipeline
#!/bin/bash
# scan-pipeline.sh
TARGET=$1
# Network scan
docker run --rm --cap-add=NET_RAW nmap-mcp \
-A $TARGET > nmap-results.txt
# Web fingerprinting
docker run --rm whatweb-mcp $TARGET > whatweb-results.txt
# Vulnerability scan
docker run --rm nuclei-mcp -u $TARGET -severity high,critical \
> nuclei-results.txtIntegration with Existing Tools
# Export Trivy results to JSON
docker run --rm -v /var/run/docker.sock:/var/run/docker.sock \
trivy-mcp image nginx:latest -f json > trivy-report.json
# Parse and filter with jq
cat trivy-report.json | jq '.Results[] | select(.Vulnerabilities)'References
Related skills
FAQ
How many security tools does mcp-security-hub include?
mcp-security-hub deploys 38 Dockerized MCP servers exposing offensive security tools including Nmap, Nuclei, Ghidra, SQLMap, Radare2, and Gitleaks for agent-driven assessments.
How does mcp-security-hub run security scans?
mcp-security-hub packages each tool as a Dockerized MCP server so Claude invokes scans and analysis through Model Context Protocol calls instead of manual CLI orchestration.