De-Anonymizing Dark Web Threat Actors: Inside Jane Intelligence Engine

Cybercriminals and illicit syndicate operators routinely operate under the illusion of total anonymity behind the Tor network, bulletproof hosting providers, and encrypted chat channels. Yet even sophisticated threat actors make subtle operational security (OPSEC) mistakes: certificate reuse across infrastructure, leaked uptime monitoring webhooks, timestamp correlation anomalies, and clearnet bridge exposures.

We built the Jane Intelligence Engine to automate threat infrastructure mapping, network telemetry correlation, and persona de-anonymization.

During our active investigation runs, Jane traced and unmasked 10 to 12 Command & Control (C2) and proxy servers distributed across Russia, Switzerland, and two hosting clusters in the United States, linking disparate infrastructure back to specific threat actor emails, phone numbers, and Telegram coordinates.

Here is an engineering overview of how the intelligence pipeline works.

The Core Problem: The Clearnet Bridge Gap

Threat actors rarely stay completely inside onion routing services. To manage payments, monitor server health, and coordinate attacks, they must interface with clearnet APIs:

• Telegram Bot APIs for alert dispatching.
• Third-party DNS and reverse-proxy services.
• Web server banners, TLS certificate serial numbers, and SSH host key fingerprints.

Jane continuously correlates these operational breadcrumbs to identify the underlying origin servers hidden behind Tor proxies.

+-----------------------------------------------------------+
|                 Threat Actor Infrastructure               |
+-----------------------------------------------------------+
       │                                     │
       ▼                                     ▼
[ Onion Routing Relay ]            [ Clearnet Bridge Misconfigs ]
       │                                - SSL Cert Serial Numbers
       │                                - SSH Host Key Hashes
       │                                - Timezone / Locale Headers
       ▼                                     ▼
+-----------------------------------------------------------+
|                 Jane Telemetry Collector                  |
|  - Custom Hardened VM Operating Environment               |
|  - Non-Attributable Packet Inspection Engine             |
|  - Cross-Source OSINT Correlation Matrix                  |
+-----------------------------------------------------------+
                               │
                               ▼
               Identified Origin Nodes:
               - 6x Russian Federation C2 Nodes
               - 3x Swiss Bulletproof Hosts
               - 2x US Cloud Instances
               - Recovered Actor Contact Profiles

1. Network Telemetry & Fingerprint Extraction

When analyzing a suspicious hidden service, Jane extracts cryptographic and network-layer markers that persist regardless of domain cycling.

// Jane Network Fingerprint Profile Builder
export interface HostFingerprint {
  ipAddress: string;
  asn: number;
  countryCode: string;
  sshKeyFingerprint?: string;
  tlsCertSerial?: string;
  jarmHash?: string;
  httpHeaderSignature: string;
}

export function generateCorrelationKey(fp: HostFingerprint): string {
  // Hash together persistent structural indicators
  const seed = `${fp.sshKeyFingerprint || 'NONE'}|${fp.tlsCertSerial || 'NONE'}|${fp.jarmHash || 'NONE'}`;
  return crypto.createHash('sha256').update(seed).digest('hex');
}

By collecting JARM hashes (active TLS server fingerprinting), SSH key footprints, and HTTP response headers, Jane queries global telemetry databases to locate servers on the public IPv4 internet running the exact same software stack and configuration.

2. Uncovering the 12 C2 Nodes

In one major investigation, a ransomware affiliate was deploying automated extortion portals. While their public face was an onion domain, their administration panel made outbound calls to an automated notification bot:

1. Telegram Webhook Exposure: We captured outgoing API payloads directed at an active Telegram Bot ID.
2. Reverse Metadata Profiling: Querying Telegram account creation records and linked bot token history surfaced the developer's handle and associated recovery phone numbers.
3. Infrastructure Cluster Identification: The SSH fingerprints matched servers on unindexed subnets in:
   • Russia: 6 nodes operating as primary payload mirrors and proxy routers.
   • Switzerland: 3 bulletproof VPS instances holding database dumps.
   • United States: 2 cloud instances used for initial staging and credential harvesting.

3. Sandboxed Forensics via Custom VM Architecture

Interacting with threat infrastructure carries severe counter-intelligence risks. If an analyst connects directly or leaks browser canvas signatures, threat actors will instantly burn the server and move to new domains.

Jane operates inside a custom, automated virtual machine running a hardened Alpine Linux distribution:

• Zero Local State: Every scan spins up an ephemeral container with randomized MAC addresses, spoofed client user-agents, and localized MTU packet sizing.
• Strict Outbound Egress Rules: All external requests route through randomized multi-hop proxies, preventing threat actors from detecting that a unified intelligence engine is scanning their network.
• Memory Forensics Integration: Incoming binary payloads are captured directly in volatile RAM and analyzed without executing on host hardware.

Key Lessons from Building Jane

1. Nobody Has Perfect OPSEC: Complex software requires maintenance. Developers push updates, check logs, and link monitoring bots. Eventually, human convenience compromises technical anonymity.
2. Correlation Over Raw Volume: You don't need petabytes of raw internet logs to solve a threat intelligence case. You need precise cryptographic fingerprints—a matching SSH key or TLS certificate serial connects distant servers across continents with 100% mathematical certainty.
3. Automation Is the Ultimate Force Multiplier: Manual OSINT tracking takes days per lead. Jane turns multi-week forensic workflows into automated real-time alerts.