Main Story

Anatomy of a Red Team Operation

Anatomy of a Red Team Operation
  • PublishedJuly 27, 2022

It’s Friday, 5p.m. An HR consultant finishes work at his client’s office and takes his loaned laptop to an IT room, off the lobby, accessed by his temporary key card. After dropping his card into a mailbox marked “Contractor Access Cards,” the consultant is ready to leave. Rainwater pours down; he struggles with his umbrella while holding the door.

“Let me get that for you,” a voice says. The consultant sees a man holding the door while entering the lobby.

Meet the Red Teamer

As a red team member, I ask the questions organizations don’t — and sometimes can’t — about their readiness in preventing, detecting, and responding to cyber attacks.

For this particular job, there was one goal: Break through the client’s security, acquire a device, and access the restricted network containing high-risk intellectual property. The client was a financial entity with custom-developed trading algorithms for predicting market trends –an appealing target for financially motivated adversaries and competitors.

On that Friday I sat in my car with the toolkit for a physical break-in. From my reconnaissance, I knew the HR consultant left promptly. I approached the building as he came into view; the moment he stopped to open his umbrella, I knew I was in.

Intrusion: Every Lock Has a Key Problem

My bag across my shoulder, I approached the key card mailbox. It was nothing special, available from any standard retailer, making replica keys easy to obtain. Opening it with a lock-pick tool was even easier. I found that the mailbox was full of cards still active for the remainder of the day. I pocketed these, took my laptop from my bag, and crossed to the IT room. The first card unlocked the door.

I noticed the consultant had left his laptop closest to the door. I’d done my research, watching employees come and go with their laptops, studying potential weaknesses those models might have. I’d watched some of the corporate videos and identified two laptop models susceptible to weaknesses I knew. The consultant’s laptop went into my bag, along with a second model. I exited the room, returned the cards, and left.

Next, I updated the client’s white team; keeping them informed is a critical responsibility. Red teams should be authentic, but they can’t succeed if they’re unsafe. It’s not about causing disruption. It’s about collaboration, communication, and education.

The Laptop Whisperer

Sitting at my desk, I opened the back panel on the consultant’s laptop to find the Trusted Platform Module (TPM) chip. This would be my first angle of attack.

For this, I used a credit card-size “logic sniffer” wired to the BIOS chip. For most modern laptops, the TPM and BIOS chips share a communication channel. The sniffer records activity passing across the BIOS chip from the motherboard and digests it for analysis on a second device. Once the key was located, I disconnected the device to begin the decryption key recovery process.

Next, I installed a backdoor on the main machine so I could discreetly maintain persistence and simulate a real-life attacker. This code would run as the laptop connected to the client’s VPN, exploiting the fact that remote workers are automatically connected after logging on. The connection provided me with a link to the network. Writing what I wanted to the disk, I had administrator-level access to the laptop. It was time to find my target.

When Convenience Kills

I found an HR application I suspected the consultant would use, and he had saved his password to autofill. It wasn’t a guarantee, but my impression was that he liked efficiency; it seemed possible that he reused passwords, so recovering one might unlock multiple doors.

I loaded the application on a second machine to use my virtualized setup. As the application read the cached password and decoded it in memory, I paused the process, freezing the data flow. The client had a legacy eight-character password policy, so I concentrated on lines of data that were  eight characters or greater. A password emerged.

After reviewing the applications available, I logged into one with the password. Finding file interaction functionality, I navigated out and found a command prompt. Now I could roam the application data of active users and software.

Software? It’s Complicated

I was now “under the floorboards” of the virtualized environment and identified six users with access to applications that stored temporary files in a location that I also had access to. These could be abused using DLL side-loading. I used this to sprinkle backdoored software libraries in these locations.

After a few minutes, I could see that one of the backdoored utilities was being run by an employee with access to the target application and data.

In total, I’d been at my desk for about 48 hours. I took screenshots and collected everything I needed to help make the necessary changes. I found the latest versions of what I was supposed to steal and exfiltrated source code files, copies of the development environment, and key assets.

Wrapping Up

The team reconvened for a post-mortem and to reflect on the indicators of compromise the security team could have been monitoring for. We offered pragmatic advice with short- and long-term defensive measures, allowing time to secure the resources for tackling the root cause and potential mitigation paths for detecting and containing similar attacks.

I detailed the attack scenarios, my general approach to the attack narrative, paths taken, obstacles observed, how they were circumvented, and how each attack was performed and structured. An analysis of the other attacks I’d prepared followed, plus an overview of remaining attack artifacts, the data accessed and where, how the data was kept safe and secure, and how anonymity was upheld.

Final Thoughts

The outcome of red teaming is never “pass” or “fail.” It’s a stress test, designed to highlight the control across the organization and how quickly attacks can be mitigated. It’s a unique opportunity to test critical assets and efficacy of security controls, training, and processes for defending your business. The goal is to ensure that any incident is just another day, rather than a headline with long-term impact.

 

By Tom Van de Wiele, Principal Threats and Technology Researcher, WithSecure

 

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.