18 Nov 2026, Wed

Russian APT ‘Star Blizzard’ Leverages Compromised Credentials and Impersonation to Breach US Targets

A sophisticated Russia-linked advanced persistent threat (APT) group known as Star Blizzard has demonstrated a chilling evolution in its tradecraft, successfully compromising a United States organization by impersonating an employee from a different US entity that it had previously breached. This development, detailed in a new report from Google’s Threat Intelligence Group (GTIG), highlights a tactical shift toward the weaponization of stolen corporate relationships to bypass traditional security perimeters.

Star Blizzard, which operates under a variety of aliases including Callisto, Cold Relic, Cold River, UNC4057, and Gossamer Bear, has been a fixture of the global threat landscape for over a decade. The group has been formally attributed by US authorities to Russia’s Federal Security Service (FSB), the successor to the KGB. Its primary mission is the systematic collection of intelligence that serves the strategic interests of the Russian government, often involving the exfiltration of sensitive data that can then be repurposed to facilitate secondary attacks against high-value targets.

According to Google’s researchers, Star Blizzard has maintained a remarkably consistent operational tempo, proving highly effective at infiltrating government officials, defense contractors, and non-governmental organizations (NGOs) across both North America and Europe. The recent incident identified by GTIG—which the company tracks specifically under the moniker "Cold Relic"—serves as a case study in how modern threat actors are converging human-centric social engineering with sophisticated cloud identity abuse.

The Anatomy of the Initial Breach

The campaign described by Google began in late 2025, when the threat actors successfully compromised an initial US-based organization. The attackers gained a foothold by seizing control of a "super admin" account, providing them with extensive privileges from the outset. To facilitate this, the group utilized a residential proxy network positioned geographically close to the target, a common tactic used to mask malicious login attempts as legitimate traffic. Once the connection was established, the hackers utilized multi-factor authentication (MFA) push notification fatigue, bombarding the target with authentication requests until the user inadvertently approved one, thereby allowing the attackers to register a new, malicious device to the victim’s tenant.

Once inside the infrastructure, the threat actors moved quickly to consolidate their access and minimize their digital footprint. They downloaded the organization’s entire cloud storage drive, providing them with a massive repository of sensitive corporate data. Simultaneously, they accessed the administrative console, where they engaged in "fingerprinting" the organization’s security defenses and actively removing logs and other evidence of their initial intrusion. To ensure they had a comprehensive map of the organization’s personnel and potential targets, they exported full employee lists and initiated a domain-wide export of account data.

Persistence was a key priority for the group. Throughout the months they remained within the victim’s environment, Star Blizzard utilized a combination of passkeys, rogue MFA devices, and backup codes to maintain their access even when security teams attempted to remediate the compromised accounts. Beyond mere persistence, the group engaged in the aggressive exfiltration of emails and cloud-hosted documents, hunting for any information that could be leveraged for future operations.

Weaponizing Trust through Impersonation

The most alarming aspect of this campaign occurred after the security team finally succeeded in removing the threat actors from the first organization’s infrastructure. Rather than retreating, the group leveraged the intelligence they had gathered during their tenure in the first victim’s network to launch a highly targeted, second-stage attack.

Google researchers observed that the threat actors began using email accounts that impersonated employees of the now-secured first organization. By assuming these identities, the attackers were able to reach out to both internal staff members and external professional associates. Because these communications appeared to originate from a known, trusted source, the attackers were able to navigate around the typical skepticism that security-conscious professionals apply to unsolicited emails.

This social engineering campaign reached a peak when the attackers, posing as an employee of the first organization, initiated contact with a second US organization via a secure messaging application. By infiltrating the professional circles shared by the employees of both companies, the hackers established a veneer of legitimacy that proved difficult for the second victim to detect. Once rapport was established, the attackers shared phishing links carefully crafted to steal credentials and Entra ID authentication tokens.

With these stolen secrets in hand, the attackers registered a new MFA device to the second organization’s single sign-on (SSO) solution. This allowed them to bypass security controls and gain persistent access to the second victim’s environment. The result was a repeat of the previous operation: the attackers performed bulk downloads of files accessible through the compromised user’s cloud drive, effectively pivoting from one breach to the next by turning a trusted relationship into an attack vector.

The Convergence of Social Engineering and Cloud Abuse

Google’s analysis of the campaign underscores a broader, concerning trend in state-sponsored cyber espionage. The "Cold Relic" methodology demonstrates that the barrier to entry for highly sensitive data is no longer solely about exploiting zero-day vulnerabilities or technical flaws. Instead, it is about the mastery of identity. By obtaining access to strategic personnel rosters, internal calendars, and sensitive operational documents, Russian intelligence services are effectively turning legitimate business communications into highly effective, low-cost attack vectors.

The threat actors were methodical in their preparation, even going so far as to change their profile names in the secure messaging application used in the second intrusion. This indicates that the group viewed the profile as a reusable asset, likely intending to deploy the same persona in future operations against other targets within the same industry or professional network.

The implications for the broader cybersecurity landscape are significant. As organizations move more of their operations to the cloud, the "identity" of an employee becomes the most critical security perimeter. When threat actors can compromise a trusted partner and then use that access to impersonate known colleagues, traditional perimeter-based security measures—and even some standard MFA implementations—become insufficient.

Google’s report serves as a stark reminder of the tenacity of state-backed actors. Star Blizzard’s ability to maintain a consistent operational tempo, despite being identified and having its tactics exposed, suggests that these campaigns are central to the Russian government’s intelligence gathering strategy. For organizations, the primary takeaway is that the risk does not end at the corporate firewall; it extends to the entire network of partners, contractors, and professional contacts that an organization interacts with daily.

As the geopolitical climate continues to drive the demand for intelligence, groups like Star Blizzard are likely to refine these impersonation tactics further. Security experts advise that organizations should implement more robust identity verification processes, particularly for communications involving sensitive data or requests for access. Furthermore, the use of hardware-based security keys, which are resistant to traditional phishing and MFA fatigue attacks, remains a critical defense against the types of credential-harvesting campaigns that Star Blizzard has successfully deployed.

The incident highlights that the "fourth battlefield"—the digital domain—is increasingly defined by the ability of intelligence services to manipulate the human and social elements of trust. By weaponizing the internal documents and social networks of their victims, groups like Cold Relic are able to maintain a persistent threat that is difficult to detect and even harder to fully eradicate, signaling a long-term challenge for the global community.