This rule detects the presence of known indicators of compromise associated with the Unknown Stealer malware, a tool frequently used by adversaries to exfiltrate sensitive data and credentials from compromised endpoints. Proactively hunting for these IOCs in Azure Sentinel allows the SOC team to identify active infections early, preventing lateral movement and data exfiltration before the threat actor achieves their objectives.
Malware Family: Unknown Stealer Total IOCs: 6 IOC Types: domain
| Type | Value | Threat Type | First Seen | Confidence |
|---|---|---|---|---|
| domain | sdn-cloudflare-js-css.cfd | payload_delivery | 2026-09-25 | 100% |
| domain | sdn-cloudflare-js.cfd | payload_delivery | 2026-09-25 | 100% |
| domain | fontawesome-cdn.cfd | payload_delivery | 2026-09-25 | 100% |
| domain | cdn-js-conhost.icu | payload_delivery | 2026-09-25 | 100% |
| domain | whitecollarmusic.com | payload_delivery | 2026-09-25 | 75% |
| domain | leeandyoondental.com | payload_delivery | 2026-09-25 | 75% |
// Hunt for DNS queries to known malicious domains
// Source: ThreatFox - Unknown Stealer
let malicious_domains = dynamic(["sdn-cloudflare-js-css.cfd", "sdn-cloudflare-js.cfd", "fontawesome-cdn.cfd", "cdn-js-conhost.icu", "whitecollarmusic.com", "leeandyoondental.com"]);
DnsEvents
| where Name has_any (malicious_domains)
| project TimeGenerated, Computer, Name, IPAddresses, QueryType
| order by TimeGenerated desc
| Sentinel Table | Notes |
|---|---|
DnsEvents | Ensure this data connector is enabled |
Scenario: A DevOps team deploys a new container image or microservice that includes a custom build of a popular open-source library (e.g., a specific version of libcurl or a patched openssl) which shares a binary hash or memory signature with one of the Unknown Stealer IOCs due to similar compilation flags or shared code segments.
dockerd, containerd-shim, crio) or where the image path matches known container storage paths (e.g., /var/lib/docker/, /var/lib/kubelet/pods/). Additionally, create an exclusion for specific executable names like node, python, or java if the IOC is a generic library hash.Scenario: An enterprise endpoint protection agent (e.g., CrowdStrike Falcon, Carbon Black, or SentinelOne) performs a full-system scan or a targeted threat hunt, loading the Unknown Stealer’s known memory signatures or file hashes into RAM for comparison, which triggers the memory-based IOC detection.
Falcon Sensor, cb.exe, SentinelOne Agent) or where the process name matches the EDR daemon. You can also filter by excluding processes that have a high thread count or are running with elevated privileges (SYSTEM/Administrator) during scheduled maintenance windows.Scenario: A legacy application or a specific version of a commercial software suite (e.g., older versions of Adobe Creative Cloud, Oracle Java, or certain .NET Framework components) uses a shared DLL or binary component that has a known collision with one of the Unknown Stealer’s file hashes due to outdated build configurations.