As reported by SecurityAffairs, Hunt.io has uncovered a sophisticated credential-theft operation that compromised 603 Windows PCs and exposed sensitive bank, government, and password manager data. The operation, attributed to Russian-speaking operators, leveraged three custom-built tools — HavoX, Phantom53, and Kraken Bridge — to maintain persistent access and exfiltrate credentials at scale. What makes this particularly notable is not the volume of victims, but the technical sophistication with which the operators defeated modern endpoint security controls.
Why This Operation Is Different
Most credential-stealing campaigns rely on commodity infostealers like RedLine or LummaC2, which primarily grab browser cookie databases and autofill data. This operation went significantly further. The operators developed custom scripts to bypass Chrome's app-bound encryption using hardcoded AES-GCM keys for Chrome versions 20 and 21, decrypt LastPass vaults by targeting AES-CBC storage, and crack 1Password master passwords through PBKDF2/HKDF/AES-GCM keyset attacks. They also lifted a live KeePass database directly from a developer's GoLand workstation — a particularly concerning vector that suggests the operators were specifically targeting development environments.
This level of tooling investment signals a shift from opportunistic theft to targeted, precision-driven credential harvesting. The operators weren't just grabbing data — they were building a pipeline to process and extract maximum value from each compromised host.
The Password Manager Reality Check
The fact that LastPass, 1Password, and KeePass all fell to different attack vectors should dispel any illusion that password managers provide unconditional protection when the endpoint is compromised.
Password managers remain the right architectural choice for credential management, but this operation underscores a fundamental truth: the security of a vault is only as strong as the endpoint it runs on. If an attacker has persistent access to a Windows machine with administrative privileges, they can intercept master passwords at entry, extract vault files, or scrape memory while the application is unlocked.
Who Is Most at Risk
Broader Implications
The operation's use of a Netherlands-based VPS with an exposed open directory suggests the operators may be a smaller, less disciplined group — a more sophisticated adversary would likely not leave their loot store publicly accessible. However, the custom tooling and the speed of deployment (1.2 million connections in three weeks) indicate this is not a novice operation either. This middle-tier capability is arguably the most dangerous segment of the threat landscape: sophisticated enough to bypass modern encryption, but agile enough to iterate quickly.
The targeting of federal identity gateways (ID.me, Login.gov) also raises the specter of downstream government system compromise. While no agency was directly breached, stolen credentials for these platforms could enable fraud, benefits theft, and lateral access to interconnected state and federal systems.
Shield53 Recommendations
Immediate Actions
- Audit endpoint security — ensure EDR/EDR solutions are deployed and actively monitoring for credential access behaviors, especially Python script execution and unusual outbound C2 connections
- Review browser credential storage policies — consider disabling Chrome/Edge password saving for high-value accounts and enforce enterprise-managed password policies
- Implement password manager hardening — enforce auto-lock timers (2-5 minutes), require re-authentication for vault access, and audit which devices have vault access
- Monitor for indicators of compromise — look for outbound connections to Netherlands-based VPS infrastructure, unusual Python process execution, and large outbound data transfers from endpoints handling sensitive credentials
Strategic Defenses
- Adopt FIDO2/WebAuthn authentication for all high-value accounts — hardware keys cannot be stolen via credential theft
- Implement conditional access policies that require device compliance checks before granting access to sensitive resources
- Segment developer environments — the KeePass theft from a GoLand workstation highlights the need for privileged access management on dev machines
- Brief employees on federal identity protection — anyone using ID.me or Login.gov should be educated on phishing indicators and credential hygiene
The most important takeaway: endpoint compromise remains the single highest-risk scenario for credential security. No encryption scheme, vault architecture, or browser hardening can fully compensate for an attacker with persistent access to a user's machine. Defense-in-depth must start at the endpoint.