Defending the Identity Control Plane: Okta Routing Rule Abuse
Scattered Spider proved that hijacking an Okta routing rule is infinitely quieter than fighting endpoint detection. Here is how to catch the control plane drift.

The blast radius of an identity plane compromise dwarfs a traditional endpoint infection. When a threat actor owns the infrastructure that issues SAML tokens and OIDC claims, they implicitly own every downstream application trusting those tokens. AWS access, financial ERPs, source code repositories, and email environments all fall like dominoes. The initial access vector was almost insultingly simple. A fluent English speaker calls the IT helpdesk, claims to have dropped their corporate smartphone in a puddle, and asks for an emergency MFA reset. Ten minutes later, the attacker has a temporary bypass code.
This is the hallmark of the threat actor known as Scattered Spider, but the real brilliance of their campaign was never the initial vishing. The brilliance was what they did five minutes after gaining access to the Okta tenant. They did not immediately dump data or deploy ransomware. They altered the authentication architecture itself, ensuring their presence would survive standard incident response evictions. By manipulating the fundamental logic that dictates how identities are verified, they created a persistent, highly privileged backdoor that evaded standard security operations monitoring.
Deconstructing the Pivot to Identity Provider Admin
Once an attacker achieves initial access via social engineering, the clock starts ticking. EDR agents are watching the endpoint. Network sensors are watching the wire. But the identity control plane is often a massive blind spot for incident responders. Scattered Spider recognized this asymmetry and weaponized it, moving laterally not through internal subnets, but through administrative privileges within the identity provider itself.
The Privilege Escalation Race
Upon compromising a standard user account, the attackers immediately scoured internal wikis, SharePoint drives, and Slack channels for exposed secrets or credentials belonging to Okta administrators. In several reported incidents, they successfully hijacked active session tokens belonging to IT staff or uncovered hardcoded service account credentials that granted them Super Administrator or Organization Administrator rights.
With administrative access secured, the threat actors effectively owned the keys to the kingdom. But maintaining that access requires extreme stealth. A loud attacker creates new administrative accounts that trigger immediate SOC alerts. A highly capable threat actor modifies the existing authentication rules to create an invisible persistence mechanism that relies entirely on legitimate system features.
Manipulating Okta Network Zones and Sign-On Policies
The core objective of this persistence mechanism is to bypass Multi-Factor Authentication for any future access, without triggering alerts for failed MFA attempts. To achieve this, the attackers exploit the native federation features of the identity provider. Specifically, they execute a masterclass in okta routing rule abuse to redirect authentication flows entirely outside of the victim organization's control.
First, the threat actor provisions their own, external Okta tenant. This rogue tenant is completely under their administrative control and is intentionally configured to approve any authentication request without ever prompting the user for a second factor. It is essentially an authentication rubber stamp controlled by the adversary.
Next, they log into the victim company's Okta administrative console and add their rogue tenant as a trusted inbound Identity Provider. This critical configuration step tells the victim tenant to explicitly trust authentication assertions coming from the attacker's infrastructure. It establishes a federated trust relationship between the victim and the threat actor.
Hijacking the Routing Logic
The final step is modifying the Identity Provider routing rules. Okta uses routing rules to determine which identity provider should authenticate a specific user based on context like network zones, device trust, or group membership. The attacker creates a new routing rule that intercepts authentication requests for their compromised administrative accounts and silently redirects those requests to the rogue tenant.
When the attacker attempts to log in the next day, the victim Okta tenant evaluates the routing rule, forwards the request to the rogue tenant, and receives a successful authentication assertion. The victim tenant logs the user in immediately. MFA is completely bypassed because the rogue tenant asserted that authentication was successful, and the routing rule mandated that the victim tenant accept that assertion blindly.
Why Standard Anomalous Login Detections Miss the Drift
Most security operations centers build their identity detection engineering entirely around the user. They write correlation rules for impossible travel, multiple failed login attempts, or anomalous logins from known bad IP addresses. These user behavior analytics form the backbone of modern identity threat detection. But routing rule modifications are an entirely different class of activity.
The Limitations of Behavior Analytics
When an attacker utilizes a rogue inbound Identity Provider via a hijacked routing rule, the authentication logs in the victim tenant look incredibly clean. The login is technically successful. There are no failed MFA prompts to trigger brute-force alerts. The source IP might even be masked or routed through a commercial VPN that overlaps with a legitimate remote worker subnet, neutralizing geography-based detections.
The blind spot exists because SOC teams monitor the data plane of authentication rather than the control plane. They watch the cars driving on the highway, but they completely ignore the construction crew rearranging the highway signs. If your detection logic only evaluates the outcome of successful logins, you will never see the architectural change that made the fraudulent login possible in the first place.
Hunting the Threat in Okta System Logs
Catching this attack requires a pivot from monitoring user behavior to monitoring administrative configuration drift. The Okta System Log provides a highly granular record of every change made to the tenant architecture. You are looking for a very specific sequence of events that indicates a federation backdoor is being installed by a rogue administrator.
The primary event type to monitor is system.routing_rule.modify. This event fires whenever an administrator alters the logic that directs authentication traffic. In isolation, a routing rule change might be a scheduled IT maintenance task. In the context of a compromised tenant, it is the smoking gun of persistent access.
eventType eq 'system.routing_rule.modify' OR eventType eq 'system.idp.lifecycle.create' OR eventType eq 'zone.create' — Okta System Log Query for Federation Abuse
Building the Attack Chain Timeline
Defenders must correlate these control plane modifications with the creation of new network zones. Attackers often create a new Okta Network Zone containing their operational IP addresses, and then tie the malicious routing rule specifically to that zone to limit exposure. The sequence typically looks like a zone.create event, followed immediately by a system.idp.lifecycle.create event to register the rogue tenant, culminating in a system.routing_rule.modify event to activate the bypass.
Security analysts triaging these alerts must parse the raw JSON payload of the system log. The target field within the log entry will reveal the exact name of the routing rule being modified, and the outcome.result field will confirm if the modification was successful. Any modification performed by an account outside of a documented change control window must be treated as a severe security incident.
Detection Controls for Untrusted Policy Changes
Identity architecture is not supposed to be dynamic. Once a company configures its inbound Identity Providers and routing rules, those configurations should remain static for months or years at a time. This inherent static nature makes it relatively easy to build a high-fidelity, low-noise detection rule that alerts security teams to control plane drift.
Immutable Identity Infrastructure
Your SIEM should trigger a critical severity alert any time an inbound Identity Provider is created or modified. There is zero legitimate reason for a helpdesk technician or even a standard systems administrator to be adding a new IdP on a Tuesday afternoon without a formal, documented engineering ticket. High-maturity organizations manage their Okta infrastructure entirely through Infrastructure as Code frameworks like Terraform, meaning any manual change in the UI is by definition an anomaly.
Furthermore, alert logic must cover any modification to Okta sign-on policies that downgrade MFA requirements. Track the policy.rule.modify event and parse the target payload. If the new rule removes the requirement for phishing-resistant MFA or allows single-factor authentication from a newly created network zone, the SOC must intervene immediately. Automated playbooks should be configured to suspend the offending administrator account and revoke all active session tokens pending an immediate investigation.
The Strategic Shift in Identity Defense
The Scattered Spider campaigns forced the security industry to acknowledge an uncomfortable truth about modern infrastructure. The identity provider is not just a centralized authentication gateway. It is a highly privileged, globally accessible administrative surface that commands your entire corporate estate. Defending it requires treating configuration drift with the exact same urgency as a triggered malware alert on a domain controller.
Investigating these control plane modifications requires rapid ingestion and parsing of identity logs before the attacker can purge them or establish deeper persistence. Whether you are writing custom Sigma rules for your SIEM or relying on MailSleuth.AI to automatically correlate system log anomalies with email forensic data, the operational mandate is completely clear. You must proactively secure the architecture that secures your users, because the adversaries are no longer bothering to hack the endpoints. They are simply logging in.
The takeaway
The Scattered Spider campaigns forced the security industry to acknowledge an uncomfortable truth about modern infrastructure. The identity provider is not just a centralized authentication gateway. It is a highly privileged, globally accessible administrative surface that commands your entire corporate estate. Defending it requires treating configuration drift with the exact same urgency as a triggered malware alert on a domain controller.
Investigating these control plane modifications requires rapid ingestion and parsing of identity logs before the attacker can purge them or establish deeper persistence. Whether you are writing custom Sigma rules for your SIEM or relying on MailSleuth.AI to automatically correlate system log anomalies with threat telemetry, the operational mandate is completely clear. You must proactively secure the architecture that secures your users, because the adversaries are no longer bothering to hack the endpoints. They are simply logging in.
We dissect phishing campaigns and email infrastructure so you don't have to.


