Adversary-in-the-Middle: How Attackers Bypass MFA With Hijacking
Multi-factor authentication is treated as the finish line. Turn it on, check the box, move on to the next project. For years, that was close enough to true. Then attackers adapted.
A modern adversary-in-the-middle (AiTM) attack doesn’t need to beat your MFA prompt. It needs your session: the proof that you already logged in and already satisfied MFA. The attacker steals that proof in real time while you’re the one typing.
That’s what makes AiTM attacks different from the credential-stuffing and brute-force attempts MFA was built to stop. An AiTM attack can turn a fully protected, MFA-enabled account into an open door without ever triggering another MFA prompt for the attacker.
This article covers what an AiTM attack is, how session hijacking bypasses MFA step by step, why traditional security tools miss it, and what detection and defense look like when the login itself isn’t the problem.
What Is an Adversary-in-the-Middle (AiTM) Attack?
An adversary-in-the-middle attack is a phishing technique where an attacker secretly relays traffic between a user and a legitimate website in real time, capturing both the user’s credentials and the authenticated session token the moment it is issued.
In plain terms, the attacker’s infrastructure sits between the user and the real service and acts as a proxy. The user enters a username and password into what looks like the legitimate login page. The page forwards those credentials to the real service in real time.
The user completes MFA normally, whether through a push notification, SMS code, authenticator app, or another supported method. The legitimate service approves the login, and the attacker, sitting in the middle of the exchange, captures the authenticated session the service just issued.
AiTM is a direct descendant of the older man-in-the-middle (MITM) attack, where an attacker intercepts traffic between two parties to eavesdrop on or alter data in transit. AiTM narrows that same positioning to a higher-value target: credentials and the live, MFA-satisfied session token that follows them.
How AiTM Bypasses MFA: Step by Step
An AiTM attack follows a consistent four-step chain. Each step looks ordinary on its own, and that is the point.
1. The Phishing Lure
The user receives a phishing email or message containing a link to what appears to be a familiar login page, such as Microsoft 365, Google Workspace, a bank, or a SaaS application they use every day.
The link routes to attacker-controlled infrastructure designed to look identical to the legitimate service.
2. The Reverse Proxy
The fake page is not simply a static copy. It functions as a reverse proxy that relays every field, click, and request to the real website in real time.
Because the user is interacting with the actual service behind the scenes, the login behaves like the genuine experience, including the MFA step. The user completes MFA normally because the real service is genuinely asking for it.
3. Session Token Theft
Once MFA succeeds, the real service issues a session token: a cookie or credential that tells subsequent requests that the user has already proved who they are and does not need to authenticate again.
Because the attacker’s proxy sits in the middle of the exchange, it captures that token as soon as it is issued.
4. Pass-the-Cookie Access
The attacker replays the stolen session token to access the account directly, with no new login and no new MFA prompt.
As far as the service is concerned, it is still the same authenticated session because technically, it is.
Why MFA Alone Can’t Stop AiTM
Here’s the reframe security teams need to internalize: MFA validates the login event. It says nothing about what happens to the session afterward.
A stolen session token is already past that checkpoint. MFA did its job, and the attacker benefited from it anyway.
This is why common forms of MFA do not stop this attack by themselves. Push notifications, SMS codes, and one-time passcodes are relayed through the attacker’s reverse proxy in real time alongside the username and password.
The user is not skipping MFA or necessarily approving an obviously fraudulent prompt. They are completing a real MFA challenge for a real login, while the attacker has a front-row seat to the entire exchange.
Phishing-Resistant MFA Raises the Bar
There is one mitigation worth emphasizing: phishing-resistant MFA, including FIDO2 security keys and passkeys, raises the bar significantly because authentication is cryptographically bound to the legitimate domain and cannot be relayed through a proxy the way a push notification or OTP code can.
Adoption is still partial across many organizations, and even where phishing-resistant MFA is deployed, it is not a complete answer by itself. It closes the initial phishing vector, but it does not eliminate the broader need to detect and respond to a session that is already in an attacker’s hands.
Why Traditional Security Misses AiTM
The blind spot comes down to one fact: the attacker is using a valid token and a legitimate session, so the access looks authorized because technically, it is.
The Login Looks Successful
MFA passed. The session was issued by the real identity provider through the real authentication flow.
Authentication logs show exactly what they would show for a normal successful sign-in because that is what actually happened.
The Access Looks Authorized
The stolen token carries the real user’s permissions, group memberships, and access rights.
Every action the attacker takes afterward, including reading email, creating inbox rules, or accessing files, looks like the user doing their job because the system has no inherent way to know that the person behind the session has changed.
Static Rules Don’t Catch It
Point-in-time checks and rule-based alerting are built to catch events such as a bad login, wrong password, failed MFA, or blocked IP.
A replayed session token can move past those controls because there may be no second login event to flag.
Catching it requires behavioral context: impossible travel between the original login and later session use, anomalous activity patterns, access from unfamiliar devices or locations, or behavior that drifts away from the user’s established baseline.
That leaves one question static, rule-based tooling cannot answer on its own: is this legitimate session still being used by the legitimate user?
Answering that requires watching how a session behaves after authentication, not only whether authentication succeeded. That is the core job of behavioral, identity-layer detection such as Blackpoint ITDR.
The MSP & Cloud Identity Angle
For MSPs and cloud-first organizations, this is not an abstract risk. Identity is the control plane.
One hijacked Microsoft 365 or Google Workspace session can lead directly to business email compromise, malicious inbox rules, payment fraud, and lateral movement into other connected systems, all without the attacker ever touching an endpoint.
The scale of the problem is already visible in Blackpoint’s own SOC data. AiTM phishing accounted for approximately 16% of cloud account disables documented in Blackpoint’s 2026 Annual Threat Report.
In those incidents, MFA functioned as designed. Attackers did not defeat authentication itself. They captured the authenticated session token issued after successful MFA and reused it, and from the cloud platform’s perspective, the resulting activity looked like a legitimate authenticated session.
For MSPs managing multiple client tenants, the stakes compound quickly. One hijacked session is not only one compromised account. It is a foothold into a client environment and potentially a path into other systems connected through shared infrastructure, applications, or trust relationships.
How to Detect & Defend Against AiTM
Effective AiTM defense is not one control. It is a layered posture across the login, the session, and the behavior that follows.
Strengthen the Front Door
Deploy phishing-resistant MFA using FIDO2 security keys or passkeys wherever possible. Prioritize high-value accounts first, including administrators, finance users, executives, and anyone with broad mailbox or tenant access.
Shorten & Protect Sessions
Tighten session lifetimes so a stolen token has a smaller window of usefulness.
Where supported, bind sessions to device and context and enable conditional access policies that factor in location, device compliance, and risk signals.
Watch How Sessions Behave
Monitor for the behavioral signals a replayed token can produce, including:
- Impossible travel, such as a login from one location followed by activity somewhere else faster than physically possible
- Access from a new or unrecognized device
- Off-hours activity
- Unexpected mailbox behavior, such as new inbox or forwarding rules
- Unusual OAuth grants or application access
Detect & Respond at the Identity Layer
None of the controls above matter enough without the ability to act quickly when suspicious behavior becomes a confirmed threat.
A hijacked session is live the moment it is replayed, so detection needs 24/7 monitoring paired with the authority to revoke sessions, disable compromised accounts, and force password resets, not simply generate an alert that waits in a queue.
Why Prevention Alone Isn’t Enough
Every control in the previous section reduces risk. None guarantees that a session will never be stolen.
Phishing-resistant MFA cannot protect a user who has not been enrolled yet. Session-lifetime policies cannot shrink the usable window to zero without affecting usability. No front-door control is airtight against every future technique or sufficiently convincing lure.
That is why identity security cannot stop at hardening the login.
Catching the use of a stolen session requires behavioral detection and response at the identity layer, not only stronger login controls. Prevention narrows the opportunity. Detection and response catch the attacker who gets through anyway and stop the activity before the session is used for greater damage.
How Blackpoint Detects AiTM & Session Hijacking
Blackpoint’s identity threat detection and response (ITDR) is built to detect identity threats including AiTM and session hijacking, backed by a 24/7 human-led SOC that investigates and contains confirmed threats.
Response can include suspending the compromised account, revoking active sessions, and forcing a password reset rather than only generating an alert for someone else to act on later.
The detection model is built around how a session is actually used across identity and cloud identity threats, not only whether the original login event looks clean. It is also designed for multi-tenant MSP environments where one compromised identity can threaten more than one client.
When the model identifies a threat that meets its confidence threshold, it can act autonomously, containing credential-based threats in under two minutes on average and in as little as 21 seconds.
Anything below that confidence threshold routes to a human SOC analyst along with the system’s full reasoning, so fast action and human judgment remain in the loop rather than one replacing the other.
AiTM Protection Goes Beyond MFA
MFA is necessary, but it is no longer sufficient on its own. An AiTM attack can turn a successful, properly completed login into a compromised session, while the authentication logs still look like an ordinary day.
A defensible identity posture combines phishing-resistant MFA with behavioral detection and fast response at the identity layer.
For MSPs, the goal is not only to stop the phishing lure. It is to detect a hijacked session while the attacker is actively using it and contain the threat quickly rather than discovering the compromise after payment fraud, mailbox abuse, or other damage has already occurred.
Strengthen Your Identity Security With ITDR
ITDR is built into the CompassOne platform, giving MSPs a single place to detect and respond to identity-based threats like AiTM alongside endpoint and cloud security instead of bolting on a separate identity tool and hoping it talks to everything else.
See how Blackpoint detects and contains session hijacking before it becomes a larger breach.
DATE PUBLISHEDNovember 18, 2026
AUTHORBlackpoint Cyber
SHARE ON
The Cost of Waiting
What Identity Attacks Cost the MSP Stolen credentials don't trip alarms. Once an attacker is authenticated, they look like anyone else, and most tools aren't watching what happens next. Learn what closing the identity response gap looks like: how identity attacks happen, where response slows down, and the questions to ask any ITDR vendor.
*Identity-based techniques play a role in nearly 90% of incident investigations