If you work in cybersecurity, you’ve probably heard the time-honored adage about cyber attacks: “It’s not a matter of if, but when.” Perhaps a better way to think of it is this: while training, experience, and familiarity with social engineering techniques help, anyone can fall for a well-constructed ruse. Everyone – including security researchers – has a vulnerability that could make them susceptible, given the right situation, timing, and circumstances.
Cybersecurity companies aren’t immune by any means. In March 2025, a senior Sophos employee fell victim to a phishing email and entered their credentials into a fake login page, leading to a multi-factor authentication (MFA) bypass and a threat actor trying – and failing – to worm their way into our network.
We’ve published an external root cause analysis (RCA) about this incident on our Trust Center, which dives into the details – but the incident raised some interesting broader topics that we wanted to share some thoughts on.
First, it’s important to note that MFA bypasses are increasingly common. As MFA has become more widespread, threat actors have adapted, and several phishing frameworks and services now incorporate MFA bypass capabilities (another argument for the wider adoption of passkeys).
Second, we’re sharing the details of this incident not to highlight that we successfully repelled an attack – that’s our day job – but because it’s a good illustration of an end-to-end defense process, and has some interesting learning points.
Third, three things were key to our response: controls, cooperation, and culture.
Controls
Our security controls are layered, with the objective of being resilient to human failure and bypasses of earlier layers. The guiding principle behind a ‘defense-in-depth’ security policy is that when one control is bypassed, or fails, others should kick in – providing protection across as much of the cyber kill chain as possible.
As we discussed in the corresponding RCA, this incident involved multiple layers – email security, MFA, a Conditional Access Policy (CAP), device management, and account restrictions. While the threat actor bypassed some of those layers, subsequent controls were then triggered.
Source: Sophos