wifi-security-mcp-server

Wireless security intelligence for AI agents.

WiFi (WPA/WPA2/WPA3), Bluetooth (BR/EDR, BLE), RF/SDR, NFC/RFID, Cellular/5G, IoT (Zigbee, Z-Wave, MQTT, LoRaWAN), MouseJack, Air-Gap, UWB, WiFi CSI, Miracast, WiFi Direct, Drone RF, WIDS/WIPS, Compliance — unified into a single MCP server.
Your AI agent gets full-spectrum wireless security analysis on demand, powered by 9 built-in knowledge bases covering 139 individual tools.

The Problem • How It's Different • Quick Start • What The AI Can Do • Tools (34) • Knowledge Bases • Architecture • Changelog • Contributing

The Problem

Wireless security assessment is one of the most fragmented disciplines in offensive security. WiFi auditing requires one set of tools, Bluetooth another, RF/SDR a third, IoT protocols yet another — and each domain has its own attack surface, its own vulnerabilities, its own compliance requirements. Today you run aircrack-ng for WPA, use Bettercap for BLE, fire up HackRF for SDR, consult different CVE databases for each chipset, reference NIST and PCI-DSS separately, and then spend hours manually correlating findings into a coherent assessment.

Traditional wireless security workflow:
  audit WiFi networks              ->  aircrack-ng / Wireshark / Kismet
  test Bluetooth security          ->  Bettercap / btlejack / Ubertooth
  analyze RF signals               ->  HackRF / GNU Radio / GQRX
  check NFC/RFID                   ->  Proxmark3 / libnfc
  assess cellular security         ->  srsRAN / OpenBTS
  test IoT protocols               ->  Zigbee sniffer / Z-Wave tools / MQTT clients
  look up wireless CVEs            ->  NVD / vendor advisories (manual search)
  check compliance                 ->  NIST / PCI-DSS / HIPAA docs (manual)
  map to MITRE ATT&CK              ->  ATT&CK Navigator (manual mapping)
  correlate everything             ->  copy-paste into a report
  ────────────────────────────────
  Total: hours per assessment, most of it context-switching between tools

wifi-security-mcp-server gives your AI agent 34 composite tools backed by 9 knowledge bases covering 139 individual capabilities via the Model Context Protocol. The agent analyzes networks, identifies vulnerabilities, maps attacks to MITRE ATT&CK, checks compliance, and generates actionable reports — in a single conversation.

With wifi-security-mcp-server:
  You: "Audit the wireless security of our office network running WPA2-Enterprise"

  Agent: -> scan_wifi_networks {encryption: "WPA2"}
         -> analyze_wifi_security {encryption: "WPA2-Enterprise", authentication: "EAP-PEAP"}
         -> find_wifi_attacks {target_type: "WPA2-Enterprise"}
         -> detect_wireless_attack {attack_indicators: "deauth frames"}
         -> audit_wireless_compliance {standard: "PCI-DSS", network_type: "WPA2-Enterprise"}
         -> "Your WPA2-Enterprise network has 3 critical findings:
            1. EAP-PEAP without certificate pinning — vulnerable to evil twin
               with hostapd-mana (MITRE T1557.002)
            2. No 802.11w (PMF) — susceptible to deauthentication attacks
               (MITRE T1498.001)
            3. PMKID hash extractable without client interaction — offline
               cracking possible with hashcat mode 22000
            PCI-DSS 4.0: 2 controls non-compliant (4.2.1, 11.2.1).
            Recommended: migrate to WPA3-Enterprise with SAE, enable PMF,
            deploy WIDS for rogue AP detection."

How It's Different

Existing tools give you raw packet captures and protocol-specific analysis. wifi-security-mcp-server gives your AI agent the ability to reason across all wireless domains simultaneously with deep knowledge-base backing.

Quick Start

Option 1: npx (no install)

npx wifi-security-mcp-server

All 34 tools work immediately. No API keys required — core functionality is knowledge-base driven.

Option 2: Clone

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

Environment variables (optional)

# WiGLE — wardriving database for real-world WiFi network lookups
export WIGLE_API_KEY=your-key

# Kismet — live wireless intrusion detection
export KISMET_HOST=127.0.0.1      # Default: 127.0.0.1
export KISMET_PORT=2501           # Default: 2501
export KISMET_API_KEY=your-key

All API keys are optional. Without them, you still get full access to all 34 tools powered by 9 built-in knowledge bases covering WiFi attacks, Bluetooth attacks, RF/SDR, NFC/RFID, cellular, IoT, hardware databases, CVE intelligence, and compliance frameworks.

Connect to your AI agent

# With npx
claude mcp add wifi-security-mcp-server -- npx wifi-security-mcp-server

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

{
  "mcpServers": {
    "wifi-security": {
      "command": "npx",
      "args": ["-y", "wifi-security-mcp-server"],
      "env": {
        "WIGLE_API_KEY": "optional",
        "KISMET_HOST": "optional",
        "KISMET_PORT": "optional",
        "KISMET_API_KEY": "optional"
      }
    }
  }
}

Start querying

You: "What attacks are possible against a WPA3-SAE network and how do I test for them?"

That's it. The agent looks up attacks, maps them to MITRE ATT&CK, recommends tools, and provides step-by-step methodology — all from built-in knowledge bases.

What The AI Can Do

WiFi Security Assessment

You: "Analyze the security of a WPA2-PSK network with WPS enabled"

Agent: -> analyze_wifi_security {encryption: "WPA2-PSK", features: ["WPS"]}
       -> crack_wps {method: "pixie_dust"}
       -> analyze_pmkid {capture_method: "hcxdumptool"}
       -> find_wifi_attacks {target_type: "WPA2-PSK"}
       -> "Critical vulnerabilities found in WPA2-PSK + WPS configuration:
          1. WPS Pixie Dust Attack — PIN recoverable in seconds via
             offline brute-force of E-S1/E-S2 nonces (Reaver + PixieWPS).
             Gives full PSK recovery without client interaction.
          2. PMKID Attack — first packet of 4-way handshake contains
             PMKID hash, crackable offline with hashcat -m 22000.
             No deauth needed, no client needed.
          3. 4-Way Handshake Capture — EAPOL frames capturable with
             airodump-ng + deauth, then offline dictionary attack.
          Priority: Disable WPS immediately (highest risk), enforce
          strong 20+ char PSK, consider migration to WPA3-SAE."

Bluetooth & BLE Analysis

You: "What are the security risks of our BLE-based access control system?"

Agent: -> scan_bluetooth {scan_type: "BLE"}
       -> find_bluetooth_attacks {target_type: "BLE"}
       -> analyze_ble {feature: "GATT", pairing_method: "JustWorks"}
       -> "BLE access control system — 4 critical risks identified:
          1. JustWorks Pairing — no MITM protection. Attacker can
             intercept pairing with Ubertooth One (MITRE T1040).
          2. GATT Characteristic Sniffing — unlock commands sent in
             cleartext, replayable with gatttool (MITRE T1212).
          3. BLE Relay/Replay Attack — relay legitimate unlock signal
             using two devices + btlejack (MITRE T1557).
          4. MAC Address Tracking — static BLE MAC enables occupant
             tracking (MITRE T1592.002).
          Recommended: enforce LE Secure Connections (LESC) with
          Numeric Comparison, encrypt GATT characteristics, implement
          rolling nonces for replay protection."

IoT Wireless Audit

You: "Assess the wireless security of our smart building IoT deployment"

Agent: -> scan_iot_wireless {protocols: ["Zigbee", "Z-Wave", "MQTT"]}
       -> analyze_zigbee {network_key_type: "default"}
       -> find_mousejack {vendor: "Logitech"}
       -> analyze_wifi_security {encryption: "WPA2-PSK"}
       -> "Smart building IoT wireless assessment — 5 findings:
          1. Zigbee: Default trust center link key (ZigBeeAlliance09)
             allows network join interception with KillerBee (CRITICAL)
          2. MQTT: Broker on port 1883 without TLS — credentials
             and sensor data transmitted in cleartext
          3. Z-Wave: S0 security with known key exchange vulnerability
             — recommend S2 upgrade
          4. MouseJack: 3 Logitech Unifying receivers vulnerable to
             keystroke injection (CVE-2016-10761)
          5. WiFi: IoT devices on same VLAN as corporate — no
             network segmentation
          Priority remediation: MQTT TLS, Zigbee install codes,
          IoT VLAN isolation, replace vulnerable Logitech receivers."

Compliance Audit

You: "Check our hospital WiFi against HIPAA wireless requirements"

Agent: -> audit_wireless_compliance {standard: "HIPAA", network_type: "WPA2-Enterprise"}
       -> analyze_medical_wireless {device_types: ["infusion_pump", "patient_monitor"]}
       -> detect_wireless_attack {attack_indicators: "rogue_ap"}
       -> "HIPAA Wireless Compliance Audit — 6 findings:
          3 NON-COMPLIANT controls:
          - 164.312(a)(1): No wireless IDS deployed for rogue AP detection
          - 164.312(e)(1): Medical devices using WPA2-PSK (not Enterprise)
          - 164.312(e)(2): No wireless traffic encryption audit logging
          3 COMPLIANT controls:
          - 164.310(a)(1): Physical AP locations documented
          - 164.312(a)(2): 802.1X with RADIUS authentication
          - 164.312(d): Certificate-based EAP-TLS for staff devices
          Remediation plan: deploy WIDS (Kismet/AirMagnet), migrate
          medical devices to WPA2-Enterprise with MAB, enable wireless
          audit logging on WLAN controller."

Tools Reference (34 composite tools)

CLI Usage

# List all available tools
npx wifi-security-mcp-server --list

# Run any tool directly
npx wifi-security-mcp-server --tool scan_wifi_networks '{"encryption":"WPA2"}'
npx wifi-security-mcp-server --tool find_wifi_attacks '{"target_type":"WPA3-SAE"}'
npx wifi-security-mcp-server --tool analyze_ble '{"pairing_method":"JustWorks"}'
npx wifi-security-mcp-server --tool lookup_wireless_cve '{"keyword":"Broadcom WiFi"}'
npx wifi-security-mcp-server --tool audit_wireless_compliance '{"standard":"PCI-DSS"}'
npx wifi-security-mcp-server --tool analyze_zigbee '{"network_key_type":"default"}'
npx wifi-security-mcp-server --tool export_report '{"format":"json"}'

Knowledge Bases (9)

The server ships with 9 built-in knowledge bases — no API keys, no external dependencies, no rate limits. Every tool query is enriched with structured intelligence from these databases.

Architecture

src/
  index.ts                    # CLI entrypoint (--help, --list, stdio server)
  protocol/
    mcp-server.ts             # MCP server setup (stdio transport)
    tools.ts                  # Tool registry — all 34 composite tools assembled here
  types/
    index.ts                  # Shared types (ToolDef, ToolContext, ToolResult)
  knowledge/                  # 9 built-in knowledge bases
    wifi-attacks.ts           # 47 WiFi attack definitions
    bluetooth-attacks.ts      # 8 Bluetooth attack definitions
    rf-attacks.ts             # 6 RF/SDR attack definitions
    nfc-attacks.ts            # 6 NFC/RFID attack definitions
    cellular-attacks.ts       # 6 Cellular/5G attack definitions
    iot-attacks.ts            # 6 IoT attack definitions
    hardware.ts               # 12 hardware device profiles
    cve-database.ts           # 17 wireless CVEs
    compliance.ts             # 5 compliance frameworks
  wifi/                       # WiFi Core tools (8)
  bluetooth/                  # Bluetooth tools (3)
  rf/                         # RF/SDR tools (1)
  nfc/                        # NFC/RFID tools (1)
  cellular/                   # Cellular/5G tools (1)
  iot/                        # IoT Wireless tools (3)
  advanced/                   # Advanced Wireless tools (6)
  infrastructure/             # Infrastructure tools (4)
  intelligence/               # Intelligence tools (4)
  detection/                  # Detection tools (2)
  reporting/                  # Reporting tools (1)

Design decisions:

• Knowledge-base driven — Unlike API-dependent tools, core functionality runs entirely offline from 9 built-in databases. No API keys, no rate limits, no external dependencies for the primary workflow.
• 34 composite tools, 139 individual capabilities — Each tool is a composite that bundles related analyses. find_wifi_attacks doesn't just list attacks — it returns methodology, required hardware, MITRE ATT&CK mapping, detection strategies, and countermeasures.
• 2 runtime dependencies — @modelcontextprotocol/sdk and zod. Nothing else. No native modules, no binary dependencies, no platform-specific code.
• Optional live integration — WiGLE API for real-world network data, Kismet for live wireless IDS. Core analysis works without either.
• Compliance-aware — Every finding is automatically mapped to relevant compliance frameworks (PCI-DSS, HIPAA, NIST, FedRAMP, ISO 27001) when applicable.
• MITRE ATT&CK mapped — Attack techniques include ATT&CK technique IDs, tactic categories, and detection data sources for SOC integration.

Limitations

• Knowledge bases are point-in-time snapshots — new attacks/CVEs require package updates
• WiGLE integration requires a free API key for network lookups
• Kismet live WIDS requires a running Kismet instance with API access
• No active wireless transmission — the server provides intelligence and methodology, not packet injection
• Hardware database covers the most common security research devices, not exhaustive
• Compliance frameworks focus on wireless-specific controls, not full framework coverage
• macOS / Linux tested (Windows not tested)

Part of the MCP Security Suite

For authorized security testing and assessment only.
Always ensure you have proper authorization before performing wireless security testing on any network or device.
Unauthorized wireless network access or interference is illegal in most jurisdictions.

MIT License • Built by Orhan Yildirim with Bun + TypeScript