Huntress Details Microsoft 365 Initial-Access Detection Using Login Enrichment

Summary
Huntress describes combining IP intelligence, user location patterns, and conditional-access events to identify suspicious Microsoft 365 logins, while noting that detections require investigation and impossible-travel detection remains challenging.
Key points
- Huntress enriches Microsoft 365 login events with IP context, including whether an address is associated with a VPN, proxy, Tor node, botnet, or cloud provider.
- Its pattern-of-life analysis estimates users’ usual login locations, helping flag logins from uncommon locations, especially when combined with anonymized IP activity.
- Conditional-access blocks can indicate that a login passed the first authentication factor, potentially exposing stolen or guessed credentials; the article cautions that these events alone are noisy and not proof of malicious activity.
- Huntress says its detectors combine unusual locations, anonymized sources, and other event context to prioritize possible account takeovers for SOC investigation.
- The company reports that anomalous-location, VPN evasion, and credential-stuffing detectors accounted for 37%, 15%, and 11% of Microsoft 365 security incidents it reported in the prior two weeks, respectively.
- Huntress says detecting account takeovers from non-VPN sources using impossible-travel signals remains difficult because of false positives.
Article Details
- Defense Focus
- Detect potential Microsoft 365 account takeovers and credential-stuffing activity early by distinguishing anomalous logins from users’ normal access patterns.
- Detection Methods
- Enrich Microsoft 365 login IP addresses with VPN, proxy, Tor, known-botnet, cloud-provider, anonymization, and geolocation attributes using Spur data.
- Build per-user location patterns and flag logins from locations with low historical percentiles.
- Correlate low-percentile locations with anonymous VPN or proxy attributes and conditional-access blocks to identify higher-priority suspected credential-stuffing events.
- Compare successful logins from unusually uncommon locations across users and ISPs.
- Correlate login events from distant locations within a short time window, including whether either event uses a VPN or proxy; the article notes this approach can produce false positives.
- Data Sources
- Microsoft 365 user login events
- Login IP addresses and geolocation
- Spur IP-enrichment attributes
- User location history and per-location percentiles
- ISP information associated with login IP addresses
- Conditional-access events and the LogonError field, including BlockedByConditionalAccess
- Login outcomes and MFA-related failures
- Rule Types
- Microsoft 365 login-event detection rules
- Correlation rules for conditional-access blocks, anonymous sources, and low-percentile user locations
- Anomalous-location and VPN/proxy detection rules
- Impossible-travel login-event correlation
- Platforms
- Microsoft 365
- Huntress Managed ITDR
- Defensive Actions
- Implement conditional-access policies; the article recommends their use and describes requiring MFA or blocking access based on conditions such as location and device compliance.
- Investigate correlated anomalous-login alerts rather than treating VPN or anonymizer use alone as proof of malicious activity.
- Rotate credentials when evidence suggests an attacker may have obtained them, and advise users to change reused passwords on other accounts.
- Use location, IP-enrichment, and account-history context to prioritize cases for SOC investigation.
MITRE ATT&CK
T1078.004 · Cloud AccountsThe article discusses suspected account takeovers in which an attacker uses compromised or stolen credentials to log in to Microsoft 365 cloud accounts.T1090 · ProxyThe article says cybercriminals use VPNs, proxies, and anonymizers, including Tor, to hide their IP addresses during operations.T1110.004 · Credential StuffingHuntress describes detecting suspected credential stuffing through Microsoft 365 login events, including attempts blocked by conditional-access policies after the first authentication factor.
