An attacker can bring a clean-looking cloud address to a Microsoft 365 sign-in and avoid the risk signals an administrator expects to catch. Huntress says its new Adversary Infrastructure option blocks IP ranges tied to investigated attacks across eligible, opted-in Managed ITDR tenants. The feature went live November 5. Existing partners start opted out, so a tenant can have Managed ITDR and still miss this particular sign-in block.

Huntress first used this approach during the Railway campaign. Huntress says device-code phishing lures reached 344 organizations across five countries in less than three weeks after the attacker used Railway, a legitimate developer cloud service, for token harvesting. Huntress’s emergency Conditional Access block of the implicated Railway ranges stopped 305 attack attempts, with zero legitimate logins affected in its 14-day fleet-traffic validation.

An incident becomes a named location

The new loop starts with a compromise in one Huntress tenant. Its SOC identifies the hosting, relay or proxy addresses involved. Huntress says detection engineers require confirmed reports from at least two unrelated partners, hosting or datacenter space, tightly scoped addresses and zero legitimate fleet logins; a second engineer reviews the nomination. “If real users log in from an address, it doesn’t ship,” the vendor writes. New campaigns first go to Huntress canary tenants.

For each campaign, Huntress says it then creates one blocking Conditional Access policy and one named location labeled [HUNTRESS] in each eligible opted-in tenant. Microsoft’s network-assignment documentation says named locations can contain public IPv4 or IPv6 ranges, and a policy can apply to selected locations. Conditional Access evaluates after first-factor authentication: the claimed prevention point is before the attacker receives a usable session token, rather than before any credential attempt reaches Microsoft. Huntress says it checks tenant policies daily, repairs drift and updates ranges as attackers move, without altering policies the customer authored.

Every block, Huntress says, produces a Low-severity incident report naming the targeted identity, source and campaign; Huntress also rotates the credential the attacker attempted to use. Those reports matter even when a block worked: a correct password offered from hostile infrastructure still deserves an identity investigation.

One address was never enough

Huntress describes two campaigns behind the rollout. In the LSHIY LLC password spray, one IPv6 range generated 81 million login attempts. The vendor reports 78 compromised accounts at 64 organizations. It says 55 of those accounts had Conditional Access policies requiring MFA, but the attacker used the ROPC legacy authentication flow, which bypassed the MFA prompt. The campaign policy blocks sign-ins from the implicated range; restricting legacy authentication closes the ROPC route across other sources. A green MFA policy assignment alone was inadequate for those sign-ins.

In the second campaign, Huntress reports 497 successful sign-ins at 79 organizations and more than 200,000 failed attempts during a July-to-September token-replay operation. The attacker moved among six addresses in the legitimate hosting provider’s 193.26.115.0/24 range. An exact-address block would leave the next address available; Huntress’s range-level policy tracks the implicated space. Treat the six addresses as attacker-used and the hosting range as shared infrastructure when reviewing other sign-ins. The company describes rogue devices, removed MFA methods, hidden inbox rules and a payroll email sent outside the organization among the cleanup work at tenants hit first. An attacker who moves to another range can still attempt sign-in, so keep legacy-authentication and device-code controls in place.

Huntress also reports 5,086 distinct attacker IP addresses touching its fleet in one 14-day window. The company sells the service it is evaluating, and its campaign and block figures are its own reporting.

Match the tenant to the policy

For an existing Managed ITDR deployment, open each relevant tenant’s ITDR settings and opt in where the organization wants Huntress to manage these campaign policies. New partners onboarded from November 5 are opted in by default. The blocking path requires Entra ID P1 or P2, Security Defaults turned off, and a healthy, fully onboarded Huntress tenant. Review the eligibility state before relying on a fleet-wide setting: an opted-in tenant that cannot run Conditional Access receives detections and incident reports after sign-in instead of the pre-session block.

In the Entra admin center, inspect the [HUNTRESS] campaign policies and their named locations for eligible tenants. For a recorded block, compare the Huntress Low-severity report’s targeted identity, source and campaign with the corresponding Microsoft sign-in result and applied Conditional Access policy. Expect an unsuccessful sign-in associated with the campaign block and no issued session from that attempt. Huntress’s daily drift repair is useful, but a policy present today still needs its enabled state and assignments checked at the tenant where users sign in. For an incident report without a blocking policy, handle the identity as a post-sign-in investigation rather than a prevented attempt.

The campaign block also leaves a separate baseline decision. Microsoft recommends a Conditional Access block on device-code flow where the organization has no documented need for it: assess existing use, test in report-only mode, preserve emergency-access exclusions and then turn the policy on. Huntress describes its Managed ISPM controls for device-code flow and legacy authentication as durable posture policies, while the new Managed ITDR policy follows specific attacker ranges. Protect both the authentication flow and the known infrastructure, then verify which policy actually denied each suspicious sign-in.