Skip to content

Latest commit

 

History

History
113 lines (84 loc) · 3.18 KB

MONITORING.md

File metadata and controls

113 lines (84 loc) · 3.18 KB

🔍 ReconMaster Monitoring System v4.0.0-Titan

High-performance, automated reconnaissance monitoring with change detection and multi-channel alerting, powered by the ReconMaster v4.0.0-Titan Asynchronous Core.

🚀 Quick Start

1. Verification

Ensure you have the latest ReconMaster dependencies installed:

pip install schedule pyyaml requests aiohttp

2. Configure Monitoring

Edit config/monitoring_config.yaml to add your scoped domains and notification channels:

targets:
  - example.com

schedules:
  daily:
    - example.com

alerting:
  enabled: true
  slack:
    enabled: true
    webhook_url: "https://hooks.slack.com/services/..."

3. Run Your First Scan

Single assessment:

python monitor/scheduler.py -t example.com

Start monitoring daemon:

python monitor/scheduler.py --daemon

📋 Features

✅ v4.0 Performance

  • Titan Asynchronous Engine: The monitoring system now leverages the v4.0.0-Titan async core, completing scheduled scans up to 5x faster than previous versions.
  • Nuclei & GraphQL Integration: Automated vulnerability re-scanning and GraphQL introspection on every scheduled interval.
  • SOAP Analysis: Continuous monitoring of SOAP endpoints and WSDL changes.

✅ Change Detection

  • 🆕 New subdomains discovered via combined passive/active discovery.
  • 🚨 Takeover detection using nuclei templates.
  • 🔌 Service monitoring: Detects new or closed ports instantly.
  • 🕷️ Endpoint monitoring: Tracks new URLs and JS files discovered by Katana.

✅ Multi-Channel Alerts

  • 📧 Enterprise Email (SMTP with SSL/TLS support)
  • 💬 Slack & Discord Webhooks (Optimized Markdown formatting)
  • 📊 Visual Dashboard: Real-time updates in monitor/dashboard.html.

📁 Artifact Structure (Monitoring)

ReconMaster/
├── monitor/
│   ├── scheduler.py       # Async-aware scheduler
│   ├── diff_detector.py   # High-speed result comparison
│   └── alerting.py        # Pro-grade alert manager
├── monitor_results/       # Persistent scan history
│   └── example.com/
│       ├── 20260208_130000/
│       │   ├── ALERT.txt  # Generated only if changes found
│       │   └── ... (full recon artifacts)
└── config/
    └── monitoring_config.yaml

🔧 Advanced Configuration

Scan Profiles

You can customize the intensity of monitored scans in the config:

scan_options:
  passive_only: false
  threads: 20
  timeout: 600

🔐 Security Best Practices

  1. Webhook Safety: Avoid committing monitoring_config.yaml with live webhook URLs. Use environment variables (if supported by your environment) or strict file permissions.
  2. Alert Thresholds: Start with daily scans for large targets to avoid "Alert Fatigue."
  3. Log Rotation: Monitor the logs/ directory to ensure disk space remains available for persistent monitoring results.

Quick Links:

Last Updated: March 09, 2026
Version: 4.0.0-Titan