dns-security-mcp

English | 简体中文 | 繁體中文 | 한국어 | Deutsch | Español | Français | Italiano | Dansk | 日本語 | Polski | Русский | Bosanski | العربية | Norsk | Português (Brasil) | ไทย | Türkçe | Українська | বাংলা | Ελληνικά | Tiếng Việt | हिन्दी

The most comprehensive DNS security toolkit for AI agents.

DNSSEC validation, subdomain takeover detection, DNS tunneling analysis, typosquatting, email security audit, certificate transparency monitoring, blocklist checking, infrastructure hardening — unified into a single MCP server.
103 tools. 13 categories. 3 dependencies. 100% local. No external API calls required for core functionality.

The Problem • How It's Different • Quick Start • What The AI Can Do • Tools (103) • Configuration • CLI Usage • Architecture • Contributing

The Problem

DNS is the foundation of every internet service, and it is the most overlooked attack surface. Misconfigured DNSSEC, dangling CNAME records, missing SPF/DKIM/DMARC, DNS tunneling, typosquatting domains, expiring certificates, open resolvers — the attack vectors are everywhere, but the tools to detect them are fragmented across dozens of specialized utilities, each with its own syntax, its own output format, and its own learning curve.

Traditional DNS security workflow:
  check DNS records              ->  dig / nslookup (manual parsing)
  validate DNSSEC                ->  dnsviz.net web UI + delv CLI
  audit email security           ->  mxtoolbox.com + dmarcian.com (2 web UIs)
  detect subdomain takeover      ->  subjack / nuclei + manual CNAME review
  find DNS tunneling             ->  Wireshark pcap + custom scripts
  monitor certificates           ->  crt.sh web UI + certspotter
  check typosquatting            ->  dnstwist CLI + manual review
  scan blocklists                ->  multirbl.valli.org + spamhaus lookup
  test infrastructure            ->  multiple CLI tools + web checkers
  generate compliance report     ->  copy-paste everything into a doc
  ─────────────────────────────────
  Total: 8+ tools, 5+ web UIs, hours of manual correlation

dns-security-mcp gives your AI agent 103 tools across 13 categories via the Model Context Protocol. The agent performs DNS reconnaissance, validates DNSSEC chains, audits email authentication, detects hijacking risks, identifies tunneling, monitors certificates, catches typosquatting, checks blocklists, hardens infrastructure, and generates compliance reports — all in a single conversation, all from local DNS queries.

With dns-security-mcp:
  You: "Run a full security audit on example.com"

  Agent: -> dns_lookup: 14 record types resolved, 3 nameservers found
         -> dnssec_full_audit: DNSSEC signed, chain valid, RRSIG expires in 12 days
         -> email_full_audit: SPF present (strict), DKIM found (3 selectors),
            DMARC p=reject — excellent email security
         -> hijack_subdomain_takeover: 2 dangling CNAMEs detected (staging, old-api)
         -> ct_unauthorized_detect: 1 certificate from unauthorized CA found
         -> bl_domain_score: Risk score 15/100 — clean
         -> infra_ns_diversity: 4 NS across 3 ASNs — good diversity
         -> report_generate: Full SARIF report generated

         "example.com has strong DNS and email security posture.
          Two critical findings: staging.example.com and old-api.example.com
          have dangling CNAME records pointing to decommissioned services —
          immediate subdomain takeover risk. One certificate from a non-
          authorized CA (Let's Debug) was found in CT logs. DNSSEC RRSIG
          expires in 12 days — schedule key rotation."

How It's Different

Most DNS tools give you raw records. dns-security-mcp gives your AI agent the ability to reason across the entire DNS attack surface simultaneously.

Quick Start

Option 1: npx (no install)

npx dns-security-mcp

All 103 tools work immediately. No API keys required for core functionality — everything runs via local DNS queries.

Option 2: bunx (faster)

bunx dns-security-mcp

Option 3: Clone

git clone https://github.com/badchars/dns-security-mcp.git
cd dns-security-mcp
bun install

Connect to your AI agent

# With npx
claude mcp add dns-security-mcp -- npx dns-security-mcp

# With local clone
claude mcp add dns-security-mcp -- bun run /path/to/dns-security-mcp/src/index.ts

{
  "mcpServers": {
    "dns-security": {
      "command": "npx",
      "args": ["-y", "dns-security-mcp"],
      "env": {
        "SECURITYTRAILS_API_KEY": "optional",
        "VIRUSTOTAL_API_KEY": "optional",
        "SAFEBROWSING_API_KEY": "optional",
        "DNS_RESOLVER": "optional"
      }
    }
  }
}

Start querying

You: "Check if example.com has any DNS security issues"

That's it. The agent runs reconnaissance, validates DNSSEC, checks email security, scans for hijacking risks, and more — automatically.

What The AI Can Do

Full Security Audit

You: "Run a comprehensive DNS security audit on target.com"

Agent: -> report_full_audit {domain: "target.com"}
       -> "Complete DNS security audit for target.com:
          CRITICAL: 2 dangling CNAMEs (subdomain takeover risk)
          HIGH: DNSSEC not enabled — vulnerable to cache poisoning
          HIGH: DMARC policy is 'none' — email spoofing possible
          MEDIUM: MTA-STS not configured
          LOW: 2 NS servers in same /24 subnet
          Score: 42/100 — significant improvements needed.
          Full SARIF report generated with 8 findings."

Email Spoofing Assessment

You: "Can someone spoof emails from our domain bigcorp.com?"

Agent: -> email_spoofability_score {domain: "bigcorp.com"}
       -> email_check_spf {domain: "bigcorp.com"}
       -> email_check_dmarc {domain: "bigcorp.com"}
       -> email_check_dkim {domain: "bigcorp.com"}
       -> "Spoofability score: 72/100 (HIGH RISK)
          SPF: Present but uses ~all (softfail) instead of -all
          DKIM: Only 1 selector found (google), missing others
          DMARC: Policy is p=none with no rua reporting
          Recommendation: Change SPF to -all, set DMARC p=reject,
          add DKIM for all sending services."

Subdomain Takeover Hunt

You: "Check all subdomains of startup.io for takeover risk"

Agent: -> hijack_subdomain_takeover {domain: "startup.io", use_ct: true}
       -> hijack_dangling_cname {domain: "startup.io"}
       -> hijack_dangling_ns {domain: "startup.io"}
       -> hijack_dangling_mx {domain: "startup.io"}
       -> "Found 4 subdomains at risk:
          CRITICAL: docs.startup.io -> CNAME to ReadTheDocs (unclaimed)
          CRITICAL: staging.startup.io -> CNAME to Heroku (app deleted)
          HIGH: old-blog.startup.io -> CNAME to Ghost.io (expired)
          MEDIUM: mail2.startup.io -> MX points to decommissioned server
          All 4 can be claimed by an attacker. Immediate action required."

DNS Tunneling Detection

You: "Analyze these DNS queries for tunneling: [list of suspicious queries]"

Agent: -> tunnel_full_scan {queries: [...], timestamps: [...]}
       -> "Tunneling probability: 94/100 (VERY HIGH)
          Entropy: 4.8 bits/char avg (normal is <3.5)
          Query length: 180+ chars avg (normal is <50)
          Tool signature: Matches iodine encoding pattern
          Covert channel: Beaconing detected at 30s intervals
          TXT payloads: Base64-encoded data in responses
          Verdict: Active DNS tunnel, likely iodine-based C2."

Typosquatting Detection

You: "Find typosquatting domains targeting mybank.com"

Agent: -> typo_full_scan {domain: "mybank.com", check_registered: true}
       -> "Generated 847 typosquat variants across 7 techniques.
          42 are registered and resolving:
          - myb4nk.com (homoglyph) -> resolves to 185.x.x.x (suspicious)
          - mybank.co (TLD variation) -> parked page
          - my-bank.com (combosquat) -> phishing page detected
          - mybanк.com (IDN/Cyrillic 'к') -> redirect to credential harvester
          HIGH RISK: 4 domains actively serving phishing content.
          Recommend takedown requests for mybanк.com and my-bank.com."

Certificate Transparency Monitoring

You: "Monitor certificates issued for our domain in the last 7 days"

Agent: -> ct_monitor {domain: "company.com", days: 7}
       -> ct_unauthorized_detect {domain: "company.com"}
       -> ct_expiry_monitor {domain: "company.com", days_threshold: 30}
       -> "Certificate activity for company.com (last 7 days):
          3 new certificates issued:
          - *.company.com by Let's Encrypt (expected)
          - api.company.com by DigiCert (expected)
          - company.com by GoGetSSL (UNAUTHORIZED — not in CAA)
          Expiring soon: 2 certificates expire within 30 days.
          Action: Investigate the GoGetSSL certificate — your CAA
          record only allows Let's Encrypt and DigiCert."

Tools Reference (103 tools)

Category Overview

Configuration

Environment Variables

All environment variables are optional. Core functionality (103 tools) works entirely via local DNS queries with zero external API calls.

# Optional: Custom DNS resolver (default: system resolver)
export DNS_RESOLVER=8.8.8.8

# Optional: SecurityTrails API for enriched passive DNS data
export SECURITYTRAILS_API_KEY=your-key

# Optional: VirusTotal API for domain reputation enrichment
export VIRUSTOTAL_API_KEY=your-key

# Optional: Google Safe Browsing API for phishing/malware URL checks
export SAFEBROWSING_API_KEY=your-key

Without any API keys, all 103 tools function using local DNS queries, public DNS resolvers, RDAP, crt.sh (Certificate Transparency), and DNS-based blocklists. API keys only add optional enrichment from third-party threat intelligence services.

CLI Usage

# Show help
npx dns-security-mcp --help

# List all 103 tools with descriptions
npx dns-security-mcp --list

# Run any tool directly from the command line
npx dns-security-mcp --tool dns_lookup '{"domain":"example.com","types":["A","AAAA","MX","TXT"]}'
npx dns-security-mcp --tool dnssec_validate '{"domain":"cloudflare.com"}'
npx dns-security-mcp --tool email_spoofability_score '{"domain":"google.com"}'
npx dns-security-mcp --tool hijack_subdomain_takeover '{"domain":"target.com","use_ct":true}'
npx dns-security-mcp --tool tunnel_full_scan '{"queries":["aGVsbG8.tunnel.evil.com","d29ybGQ.tunnel.evil.com"]}'
npx dns-security-mcp --tool typo_full_scan '{"domain":"mybank.com","check_registered":true}'
npx dns-security-mcp --tool report_generate '{"domain":"example.com","format":"sarif"}'

# Using Bun (faster startup)
bunx dns-security-mcp --tool bl_domain_score '{"domain":"suspicious-site.xyz"}'

# With custom DNS resolver
DNS_RESOLVER=1.1.1.1 npx dns-security-mcp --tool dns_propagation '{"domain":"example.com"}'

Architecture

src/
  index.ts                # CLI entrypoint (--help, --list, --tool, stdio server)
  protocol/
    mcp-server.ts         # MCP server setup (stdio transport)
    tools.ts              # Tool registry — all 103 tools assembled here
  types/
    index.ts              # Shared types (ToolDef, ToolContext, ToolResult)
  utils/
    dns-client.ts         # Shared DNS query client (UDP/TCP, dns-packet)
    rate-limiter.ts       # Per-provider rate limiter
    cache.ts              # TTL cache for DNS responses
  dns/                    # DNS Reconnaissance tools (12)
  dnssec/                 # DNSSEC Validation tools (8)
  email/                  # Email Security tools (9)
  hijack/                 # Hijacking & Takeover tools (9)
  tunnel/                 # DNS Tunneling Detection tools (7)
  ct/                     # Certificate Transparency tools (7)
  domain/                 # Domain Intelligence tools (10)
  typo/                   # Typosquatting & Brand Protection tools (8)
  blocklist/              # Blocklist & Reputation tools (6)
  infra/                  # Infrastructure Security tools (9)
  privacy/                # DNS Privacy tools (6)
  threat/                 # Passive DNS & Threat Intel tools (7)
  report/                 # Compliance & Reporting tools (5)
  data/                   # Static data (DKIM selectors, DNSBL lists, tool signatures)

Design decisions:

• 3 dependencies, nothing else — @modelcontextprotocol/sdk for the MCP protocol, zod for input validation, dns-packet for raw DNS packet crafting. No bloated dependency tree. No native modules. No C bindings.
• 100% local core — Every tool works via direct DNS queries using dgram (UDP) and net (TCP) sockets with dns-packet for wire-format encoding/decoding. No external API calls required for any core functionality.
• 13 providers, 1 server — Each security category is an independent module. The AI agent picks which tools to use based on the investigation context.
• Raw DNS access — Unlike tools built on dns.resolve(), dns-security-mcp crafts raw DNS packets. This enables DNSSEC record queries (DS, DNSKEY, RRSIG, NSEC), CHAOS class queries (version.bind), non-recursive queries (cache snooping), AXFR attempts (zone transfer), and EDNS options — none of which are possible with Node's built-in DNS module.
• TTL caching — DNS responses are cached respecting record TTL values. CT log queries cached for 15 minutes. RDAP responses cached for 10 minutes. Prevents redundant queries during multi-tool workflows.
• Graceful degradation — Missing API keys never crash the server. Optional enrichment tools return clear messages: "Set SECURITYTRAILS_API_KEY for enriched passive DNS data."
• Multiple output formats — Reports can be generated as JSON (machine-readable), Markdown (human-readable), or SARIF (IDE/CI integration).

Part of the MCP Security Suite

Contributing

Contributions are welcome. See CONTRIBUTING.md for guidelines.

For authorized security testing and assessment only.
Always ensure you have proper authorization before performing DNS security testing on any target.

MIT License • Built by Orhan Yildirim • TypeScript + Bun