ProBackend
Identity & Access
Jul 5, 20266 min read

When Login Isn't Enough: Credential Abuse, Infostealers, and the New Rules of Account Security

When Login Isn't Enough: Credential Abuse, Infostealers, and the New Rules of Account Security

Credential abuse drives 22% of breaches, infostealer malware surged 84% in a year, and session-token theft walks straight past MFA. Here's how the attacks actually work — and what continuous verification buys you.

The identity controls that worked in 2022 are losing

Passwords were the first problem, and nobody should be surprised by that. What should surprise you is that the fixes we bolted on top — MFA, SSO, password-reset workflows that feel securely modern — are being walked through rather than broken.

Two numbers tell the story. Verizon's 2024 Data Breach Investigations Report found that stolen credentials featured in 22% of breaches, making them the single largest initial attack vector. And across 2024, infostealer malware infections jumped 84% year-over-year, according to the 2025 Global Threat Intelligence Report. More credentials in play, and a more industrialized way of harvesting them.

If your organization manages thousands of human and non-human identities across cloud, SaaS, endpoints, and remote environments, this isn't someone else's incident. The endpoint you govern least is probably where your most durable session tokens live.

MFA fatigue and the war of attrition

The tactic is crude and it works: blast a user with push prompts until they approve one just to make it stop. Attackers who already have a stolen password don't need to trick you into typing anything clever. They just need you tired. It's 2 a.m., the phone won't stop buzzing, and somebody taps "Approve" without reading a word. The attacker is in.

The fix is not fewer factors — it's better factors. Number matching and phishing-resistant methods like FIDO2 passkeys take the human out of the approval loop, because there's nothing to rubber-stamp when the prompt requires you to type a number displayed on the device you're actually authenticating to. Until every user is on those methods, MFA fatigue belongs in your incident response playbook as a known, living threat.

Adversary-in-the-middle phishing steals the session, not just the password

Here's the part that should keep identity teams up at night. Modern phishing kits stand as a proxy between the user and the real login page. You type your password, you pass your MFA challenge — and the kit captures everything in transit, including the session token minted on the other side of authentication.

That's the mechanics behind the claim that session theft can "bypass MFA." The kit doesn't defeat the second factor. It doesn't need to. The session cookie it relays and captures lets the attacker ride an already-authenticated session, so the login page is no longer the battleground at all. From the identity provider's point of view, it looks like a user who logged in properly.

This is exactly the kind of attack our write-up on MFA-bypassing phishing toolkits targeting Microsoft 365 walks through in detail.

The endpoint: credential theft's soft underbelly

Infostealer malware like RedLine, Raccoon, and Vidar scrapes credentials out of browsers, cryptocurrency wallets, and local files. In fact, over 73 million credentials were exposed through just three infostealer families — RedLine, Vidar, and Raccoon. Then there's the more subtle problem: session cookie theft.

You don't need a password at all when you have a live cookie. Modern stealers copy session cookies directly out of browser storage, and because the session token lives outside the browser's password vault, resetting the password — the response most IT departments reach for first — does nothing to evict the attacker. The token keeps working.

How do these stealers land? Malicious browser extensions and trojanized software installations are two of the most common infection paths, both of which sail through most endpoint controls because the user technically consented to the install. And this is malware-as-a-service: crimeware platforms package infostealers with phishing toolkits, infrastructure, and support, so the operator pool is deep and low-skill. An 84% surge in a single year isn't a mystery when the barrier to entry is a subscription.

The identity inventory nobody owns

Here's the organizational awkwardness underneath all of this. IT runs the endpoint and the patch cycle. The security team builds the incident response playbook. Identity and access management sits in a third reporting line entirely. So the endpoint gets treated as an IT ticket, the session-token attack gets treated as a security alert, and the identity that connects them — a user's stolen, still-valid cookie minted through an endpoint nobody patched — gets treated by nobody.

The gaps are structural, not technical. And they're load-bearing: the moment your human identities number in the thousands, plus the non-human identities quietly accumulating in every integration, SaaS tenant, and CI runner, the seams are where attackers live.

Why continuous verification is the way out

For most of enterprise security, it went like this: authenticate, get a token, trust it until it expires. The entire architecture assumed that if you passed the door once, you belonged inside for the life of the session. Session-token theft exploits that assumption perfectly.

Continuous verification breaks the assumption. Instead of a one-time decision stretched over hours, the system re-evaluates signals throughout the session: where the traffic originates, whether the device posture still looks like the device that signed in, whether behavior matches the baseline. When a cookie gets replayed from a new ASN in another country, the session gets challenged or killed even though the credentials are technically valid. Password rotation matters less when sessions are short-lived and re-attested. The playbook flips: instead of assuming trust and trying to detect the exceptions, assume nothing and keep checking.

No single product does all of this today, and honestly, no single team owns it either — it spans identity, network, and endpoint. That's the hard part, and it's organizational before it's technical.

Where a security & compliance analyst should focus first

If you do the identity-adjacent work — incident response, access reviews, compliance reporting — you don't have to wait for a platform re-architecture to get useful. Start with the low-glamour work. Inventory first: you can't defend sessions you don't know exist. Then instrument the auth logs to flag impossible travel and session anomalies — a session that keeps working after a password reset, or a cookie that suddenly appears at an unfamiliar IP, should be an alert, not a mystery. Prioritize phishing-resistant MFA for the accounts whose sessions matter most: admins, finance, anyone with tenant-level rights. And pressure-test your assumptions: ask what happens today if a valid session token walked out the door.

There's a broader hygiene angle on this in our cyber hygiene and identity defense analysis.

What to hold onto

The controls you built assumed a world where authentication was a moment. It isn't anymore — it's a session that attackers will try to steal wholesale, via fatigue prompts, proxied login pages, and malware sitting on the laptop of someone who passed every check you designed.

So: phishing-resistant factors over push prompts. Short-lived, re-verified sessions over long-lived tokens. Treat cookie theft as credential theft, because to an attacker it's the same thing. And stop measuring identity security by logins blocked. Start measuring by how quickly a stolen session dies.

If your answer to "how fast?" is "we'd have to check the logs," congratulations: that's the project.

Article illustrationArticle illustration