Matomo (Piwik) RUM - Tech Detect

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""Detects Matomo (formerly Piwik) Web Analytics & RUM artifacts."""

import re
from kittysploit import *
from lib.protocols.http.http_client import Http_client


class Module(Scanner, Http_client):
    __info__ = {
        'name': 'Matomo (Piwik) RUM - Tech Detect',
        'description': 'Detects Matomo (formerly Piwik) Web Analytics & RUM artifacts. Identifies Standard Tracking, Image Tracking, Tag Manager, and Cloud instances using strict matching.',
        'author': ['KittySploit Team'],
        'severity': 'info',
        'tags': ['web', 'scanner', 'technology', 'tech', 'rum', 'matomo', 'piwik', 'analytics'],
        'agent': {
            'risk': 'active',
            'effects': ['network_probe'],
            'expected_requests': 1,
            'reversible': True,
            'approval_required': False,
            'produces': ['tech_hints', 'risk_signals', 'endpoints'],
            'cost': 1.0,
            'noise': 0.3,
            'value': 1.0,
            'requires': {
                'min_endpoints': 0,
                'min_params': 0,
                'tech_hints_any': [],
                'tech_hints_all': [],
                'specializations_any': [],
                'risk_signals_any': [],
                'auth_session': False,
                'capabilities_any': [],
                'capabilities_all': [],
                'confidence_min': {},
                'confidence_min_any': {},
                'endpoint_pattern_any': [],
                'param_any': [],
                'api_surface_ready': False,
            },
            'chain': {
                'produces_capabilities': [
                    {
                        'capability': 'admin_surface',
                        'from_detail': '',
                    },
                ],
                'consumes_capabilities': [],
                'option_bindings': {},
                'suggested_followups': [],
            },
        },
        'references': [
            'https://matomo.org/',
            'https://developer.matomo.org/guides/tracking-javascript-guide',
            'https://developer.matomo.org/api-reference/tracking-javascript',
            'https://matomo.org/faq/on-premise/installing-matomo/',
            'https://matomo.org/matomo-cloud/',
        ],
    }

    def run(self):
        path = '/'
        r = self.http_request(method='GET', path=path, allow_redirects=True)
        if not r or r.status_code != 200:
            return False
        body = (r.text or "").lower()
        headers = "\n".join(f"{k}: {v}" for k, v in r.headers.items()).lower()
        body_any = ('<html',)
        header_any = ('text/html',)
        body_regexes = ('window\\._paq', '_paq\\.push\\s*\\(', 'var\\s+_paq\\s*=\\s*window\\._paq', 'MatomoTagManager', 'window\\.MatomoTagManager', '_paq\\.push\\s*\\(\\s*\\[\\s*[\'"]trackPageView[\'"]', '_paq\\.push\\s*\\(\\s*\\[\\s*[\'"]setSiteId[\'"]', '_paq\\.push\\s*\\(\\s*\\[\\s*[\'"]setTrackerUrl[\'"]', '_paq\\.push\\s*\\(\\s*\\[\\s*[\'"]enableLinkTracking[\'"]', '(Matomo|Piwik)\\.getTracker\\s*\\(', 'src\\s*=\\s*[\'"][^\'"]*(matomo|piwik)\\.js[\'"]', 'g\\.src\\s*=\\s*[^;]*(matomo|piwik)\\.js', '[\'"]https?://[^\'"]+/(matomo|piwik)\\.php[\'"]', 'u\\s*\\+\\s*[\'"](matomo|piwik)\\.php[\'"]', '[a-z0-9-]+\\.matomo\\.cloud/', '<img[^>]+src\\s*=\\s*[\'"][^\'"]*(matomo|piwik)\\.php\\?idsite=', 'name=[\'"]generator[\'"]\\s+content=[\'"](Matomo|Piwik)',)
        if (any(m in body for m in body_any)) and (any(m in headers for m in header_any)) and (any(re.search(rx, body, re.I) for rx in body_regexes)):
            self.set_info(
                severity='info',
                reason='Matomo (Piwik) RUM - Tech detected',
                path=path,
            )
            return True
        return False