{
  "family": "doubleagent",
  "sample_count": 1,
  "category": "advanced_threat",
  "description": "<h3>Executive Summary</h3>\nDoubleAgent is not a traditional malware family, but rather a highly sophisticated, open-source code injection and persistence technique. It exploits a legitimate feature of the Windows operating system—the Application Verifier—to inject a malicious DLL into *any* process during its initialization. While originally documented by researchers as a proof-of-concept, the DoubleAgent technique is actively abused by advanced threat actors and APT groups to establish completely invisible persistence and bypass next-generation endpoint security controls.\n\n<h3>Technical Architecture and Exploitation</h3>\nThe Windows Application Verifier is a developer tool designed to monitor applications for memory corruption and handle leaks. It works by loading a specific \"verifier provider DLL\" into the target process as it starts.\n\nThe DoubleAgent technique abuses this mechanism:\n<ul>\n<li><strong>Registry Manipulation:</strong> The attacker modifies a specific, legitimate Windows Registry key (`HKLM\\SOFTWARE\\Microsoft\\Windows NT\\CurrentVersion\\Image File Execution Options\\<TargetProcess.exe>`). They create a subkey named `VerifierDlls` and point it to their malicious DLL payload.</li>\n<li><strong>Universal Injection:</strong> Once the registry key is set, the Windows OS itself will automatically load the attacker's malicious DLL into the target process (e.g., `explorer.exe`, `chrome.exe`, or even the antivirus engine's process like `MsMpEng.exe`) every single time that process is launched.</li>\n<li><strong>Absolute Stealth:</strong> Because the injection is performed natively by the OS loader *before* the application even begins executing its own code, traditional antivirus hooks and user-mode EDR agents are completely blind to the injection.</li>\n</ul>\n\n<h3>Threat Impact</h3>\nThe DoubleAgent technique is a critical threat. It allows an attacker to gain full control over any application on the system. By injecting into the antivirus process itself, the attacker can silently neutralize the endpoint's primary defense mechanism, rendering it incapable of detecting subsequent malicious activity.\n\n<h3>Defense and Resilience Strategies</h3>\n<ul>\n<li><strong>Registry Monitoring (FIM):</strong> Traditional file scanning cannot stop DoubleAgent. Defenses must rely on strict File Integrity Monitoring (FIM) and EDR rules that monitor the `Image File Execution Options` (IFEO) registry keys for unauthorized modifications.</li>\n<li><strong>Protected Processes:</strong> Utilize Windows Protected Process Light (PPL) technology. Processes running as PPL (which should include enterprise EDR agents) prevent unsigned or unauthorized DLLs from being loaded, even via the Application Verifier mechanism.</li>\n<li><strong>Privilege Management:</strong> The initial execution of the DoubleAgent technique requires administrative privileges to modify the HKLM registry hive. Enforcing the Principle of Least Privilege (PoLP) and strict UAC controls prevents the initial registry modification required to set up the attack.</li>\n</ul>",
  "cta": "Published by the SystemHelpdesk team.",
  "aliases": [
    "HackTool.DoubleAgent",
    "Exploit.ApplicationVerifier",
    "Riskware.DoubleAgent"
  ],
  "enrichment_level": "expert-seo",
  "faq": [
    {
      "@type": "Question",
      "name": "What is Doubleagent and what threat category does it belong to?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Doubleagent is classified as a Advanced Threat. Documented across security telemetry, it is engineered to execute unauthorized operations, establish persistent footholds on target endpoints, and facilitate malicious campaigns targeting enterprise and individual environments."
      }
    },
    {
      "@type": "Question",
      "name": "How does Doubleagent establish persistence and evade defensive controls?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Doubleagent typically establishes persistence through registry run keys, scheduled tasks, or service installations, while evading signature-based antivirus solutions through code obfuscation, memory injection, or anti-analysis sandbox checks."
      }
    },
    {
      "@type": "Question",
      "name": "What are the essential incident response steps for mitigating a Doubleagent compromise?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Security teams responding to Doubleagent should immediately isolate affected endpoints from the network, perform memory triage to capture volatile forensic artifacts, audit registry and startup persistence mechanisms, revoke exposed user credentials, and restore systems from verified, clean backups."
      }
    },
    {
      "@type": "Question",
      "name": "What lateral movement and credential access tactics does HackTool.DoubleAgent leverage in enterprise intrusions?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "HackTool.DoubleAgent frequently utilizes privileged account harvesting techniques such as LSASS memory dumping, Kerberoasting, and token impersonation to escalate privileges. Once administrative credentials are acquired, it pivots across the victim network utilizing living-off-the-land binaries (LOLBins) such as WMI, PowerShell remoting, and PsExec, or abuses remote desktop protocols (RDP) to reach critical internal repositories and domain controllers."
      }
    },
    {
      "@type": "Question",
      "name": "How does HackTool.DoubleAgent maintain covert communications and evade perimeter network monitoring?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "HackTool.DoubleAgent often relies on multi-stage command and control (C2) architectures utilizing encrypted protocols over standard web ports (HTTPS/443 or DNS tunneling) to blend in with legitimate enterprise network traffic. Many variants incorporate proxy networks, domain fronting, or cloud services (such as Microsoft Graph API, Google Drive, or Telegram) to masquerade C2 channels, bypassing traditional signature-based egress filtering."
      }
    }
  ],
  "faq_count": 5,
  "mitre_attack": [
    "T1546.012",
    "T1055.001",
    "T1562.001"
  ],
  "cisa_advisory": null,
  "last_updated": "2026-10-15",
  "type": "HackTool / Exploit Technique",
  "target_industries": [
    "Global / Opportunistic"
  ],
  "motivation": "Opportunistic",
  "threat_actors": [
    "Unknown / Cybercriminal"
  ],
  "target_geographies": [
    "Global"
  ],
  "attack_stage": [
    "execution",
    "persistence"
  ],
  "ecosystem": [
    "windows_pe"
  ],
  "is_emerging": false,
  "mitre_attack_detail": [
    {
      "id": "T1546.012",
      "name": "Event Triggered Execution: Image File Execution Options Injection (Disabling Task Manager)",
      "tactic": "Privilege Escalation"
    },
    {
      "id": "T1055.001",
      "name": "Dynamic-link Library Injection",
      "tactic": "defense-evasion"
    },
    {
      "id": "T1562.001",
      "name": "Impair Defenses: Disable or Modify Tools",
      "tactic": "Defense Evasion"
    }
  ],
  "containment_steps": [
    "Enforce forensic network isolation on all suspected endpoints while maintaining an out-of-band telemetry conduit to capture live Doubleagent adversary activity.",
    "Perform full memory dumps across identified staging nodes to extract memory-only implants, kernel rootkit hooks, and decrypted Doubleagent C2 infrastructure IP/domain lists.",
    "Review Active Directory domain controller authentication logs for anomalous Kerberos ticket requests (Pass-the-Ticket, Golden Ticket) and unauthorized DCSync operations associated with Doubleagent.",
    "Block all confirmed C2 domains, dynamic DNS endpoints, and proxy IP addresses across perimeter firewalls, proxies, and DNS security gateways.",
    "Conduct an enterprise-wide sweep across surrounding subnets for secondary webshells, WMI event subscriptions, or remote execution tooling left behind by Doubleagent operators."
  ],
  "what_to_avoid": [
    "Do not assume the incident is resolved simply because an antivirus tool removed the initial Doubleagent dropper; always verify that secondary payloads and persistence hooks have been eradicated.",
    "Avoid logging in with Domain Admin or privileged credentials on a machine suspected of running Doubleagent until memory triage is complete.",
    "Do not delay host isolation; automated lateral movement and credential dumping can occur within minutes of initial execution."
  ],
  "target_platforms": [
    "Windows 11",
    "Windows 10",
    "Windows Server 2022",
    "Windows Server 2019"
  ]
}