Detecting Real-Time Phishing Attacks
In the last two installments in our series on phishing resistance, we discussed phishing resistant authenticators and how to gather signals about phishing lures directly from your users. Now let’s drill down into detection and response: what signals does Okta’s System Log provide that are indicative of in-flight phishing campaigns?
Okta’s Defensive Cyber Operations team routinely identifies phishing infrastructure configured to imitate an Okta sign-in page and proactively notify Okta customers when suspicious infrastructure we detect appears to be targeting their users. Since March 2020, we have delivered over 1000 notifications to customers.
Proactive alerts provides customers opportunities to:
- Deny outbound requests to phishing infrastructure from managed devices,
- Request takedowns when the infringing site goes live, and
- Adjust access policies accordingly.
There are no guarantees, however, that every phishing domain will be detected in advance of a campaign. And even when they are, there is often a short window of exposure before takedowns take effect.
So the harder problem for defenders is how to quickly identify threat actor activity and remediate any exposure while users are under attack.
FastPass is your secret weapon
Okta’s passwordless solution, FastPass, offers strong resistance against real-time phishing attacks.
When your users are enrolled in FastPass, Okta can provide defenders a high-fidelity signal for when user applications are being targeted by attackers wielding real-time (AiTM) proxies.
This Early Access feature is available for self-service on Okta Identity Engine - select "Phishing Resistance for FastPass" under Settings > Features in the Admin Console.
If one or more users enrolled in FastPass is targeted using AiTM phishing kits, Okta Identity Engine identifies the failed origin check and generates a unique event in Okta System Log:
eventType eq "user.authentication.auth_via_mfa" AND result eq "FAILURE" AND outcome.reason eq "FastPass declined phishing attempt"
The utility of that single system log event can’t be understated. In many scenarios, it’s likely to be the earliest available signal about an in-flight attack, and includes key details about the phishing infrastructure used by the adversary.
Why are those details so important? As we previously discussed in this series, there are relatively few organizations today that are 100% passwordless. Even in organizations where a majority of users are protected by phishing resistant factors, there are often groups of users with little choice but to rely on authenticators that are less resistant to phishing.
In our experience, opportunistic threat actors can’t or don’t discern between what authenticators are available to any given user. They rely on harvesting or enumerating large numbers of usernames from public sources during the reconnaissance phase of an attack. If an adversary sends a phishing email to ~100 users, their lures are likely to reach targets enrolled in a broad variety of factors.
So any early detection offers opportunities to:
- Prevent other users from accessing (or authenticating via) the attacker’s infrastructure,
- Evaluate if other users were previously targeted via the same infrastructure,
- Evaluate if other users have entered credentials, and
- Evaluate if any of the phishing activity resulted in an account takeover.
Many of these actions can be automated using Okta Workflows (or using a third-party SOAR solution).
Okta Workflows can be used, for example, to:
- Extract the IP of the attacker’s proxy server.
- Assess the reputation of the IP (by checking the ratio of successful to unsuccessful authentication events from that IP over the weeks or months prior to the incident).
- Search System Log to check whether any other users successfully authenticated via a suspicious IP. If any value is returned, the flow can automatically clear the user’s sessions.
- If users entered a password as part of the authentication flow (irrespective of whether they successfully authenticated), the flow can call System Log to check whether the user’s corporate email application was accessed during the session in question. This can help determine whether to reset the user’s password.
- Raise a request to add the IP to an org-wide blocklist (network zone) to prevent future authentication requests via the attacker’s infrastructure.
To see these ideas in action, we recommend catching up on the recorded sessions delivered on FastPass phishing resistance at this week's Oktane22 conference [Registration Required]: