RC RANDOM CHAOS

Exposed Infrastructure Reveals AI-Powered Ransomware Operations

· via The Hacker News

Original source

Exposed Infrastructure Reveals How Ransomware Affiliates Used AI Agents to Execute Intrusions

The Hacker News →

Researchers identified a large-scale intrusion operation managed by INC and The Gentlemen ransomware affiliates that automates attacks through an AI framework called Phantom. The Russian-speaking operators function as both ransomware actors and infrastructure providers for wider ecosystem partners. Phantom executes intrusions through a human-in-the-loop model: operators submit victim-specific prompts to commercial AI models hosted via OpenRouter, review each action, and approve execution. When models decline requests, operators reframe them as authorized red-team assessments rather than selecting different models.

The attack chain exploits public-facing edge infrastructure including FortiClient EMS, SimpleHelp, cPanel, and VPN gateways, using multiple unpatched vulnerabilities for initial access before moving through lateral movement, persistence, and exfiltration stages leading to ransomware deployment. The attackers weaponized 18 specialized AI skill modules within Phantom for reconnaissance, reverse engineering, exploit research, and EDR fingerprinting. Supporting tools include WormGPT-MCP (32 integrated modules), polymorphic Ghost C2 variants, RevSocks proxies for network pivoting, and EDR-killing utilities. The operators maintained AI-managed knowledge bases tracking victim details, compromised credentials, and campaign progress. Credential reuse emerged as the primary force multiplier—in documented cases, a single password provided access across VPN, servers, remote management, backup, and security tools simultaneously.

The exposed infrastructure logged activity from 62 targeted organizations, with four ending up on public leak sites after extortion attempts. The attackers adopted a data-centric extortion methodology, analyzing stolen financial records to calculate ransom amounts below insurance thresholds, then escalating pressure through masked proof samples and vishing calls threatening regulatory notification. The U.S. accounted for over half the targets, concentrated in industrial, manufacturing, and energy sectors, followed by healthcare. No payment records have surfaced, leaving the campaign’s financial success unclear, though the scale and sophistication indicate a mature operational infrastructure serving multiple threat actors simultaneously.

Read the full article

Continue reading at The Hacker News →

This is an AI-generated summary. Read the original for the full story.