Cold Relic weaponizes a Google Workspace breach as a launchpad
Google’s Mandiant and Threat Intelligence Group (GTIG) uncovered a 2026 intrusion at a U.S.-based organization that traced back to the Russia‑linked espionage outfit Cold Relic, formerly known as COLDRIVER and UNC4057.
Why the cloud angle matters
Instead of a classic phishing lure, the group appears to have compromised a Google Workspace tenant and then used the foothold to move laterally into the target environment. Leveraging a legitimate cloud service masks malicious traffic and gives the attackers a trusted channel for command‑and‑control.
Technical breadcrumbs
- Initial foothold: a Google Workspace account that was either hijacked or mis‑configured.
- Post‑compromise activity: creation of unauthorized OAuth app credentials, enabling the adversary to request tokens and impersonate legitimate users.
- Persistence: deployment of custom PowerShell scripts that harvested email metadata and exfiltrated documents to external storage.
Who’s in the cross‑hairs?
The victim was a U.S. organization operating in a sector that routinely uses Google Workspace for collaboration. No public details about the industry or data impact have been released.
Unconfirmed details
Unverified: Some reports suggest the attackers may have stolen Google Workspace data before the breach, but GTIG’s public findings do not confirm that claim.
Defensive takeaways
- Audit OAuth app permissions regularly and revoke any that are not business‑critical.
- Enforce strict MFA on all Google Workspace accounts, especially privileged ones.
- Monitor for anomalous token‑generation activity via Google’s security logs.
- Segment internal networks so that a compromised cloud credential cannot reach critical on‑prem assets without additional checks.
- Engage with a trusted incident‑response partner promptly when unusual cloud activity is detected.
