Ken Huang

Cigent Pre-Boot Authentication (PBA) achieves NIAP listing and Common Criteria certification, reinforcing Cigent's leadership in protecting data-at-rest for the defense community.

24

1 Comment

A fresh wave of digital intrusions, tied to Russian operatives known as APT28, emerges through findings uncovered by S2 Grupo’s LAB52 analysts. Throughout late 2025 into early 2026, these efforts quietly unfolded across Western and Central European institutions. Dubbed Operation MacroMaze, the pattern reveals reliance on minimalistic yet precisely timed actions. Instead of complex tools, attackers favored subtle coordination - bypassing alarms by design. Each phase unfolded with restraint, avoiding flashiness while maintaining persistence behind the scenes. Starting the operation, cyber actors send targeted emails with harmful attachments designed to trick users. Instead of using typical methods, these documents include an XML feature named “INCLUDEPICTURE.” That field points to a JPG stored on webhook[.]site, acting as a hidden reference. As soon as someone views the file, the system pulls the image from that external address. Unlike passive downloads, this transfer initiates a background connection outward. Midway through loading, the request exposes details about the user’s environment automatically. So, without visible signs, attackers receive confirmation plus technical footprints tied to the access event. https://lnkd.in/gBpc8R4X

5

At the recent Central Maryland Chamber AI & Tech Summit, the concept of "Harvest now, decrypt later" was a focal point of discussion. This innovative strategy involves gathering and securely storing encrypted data, envisioning potential decryption advancements in anticipation of the forthcoming Quantum revolution. The essence lies in exploiting future decryption technologies to unlock currently inaccessible data, a concept known as Q-Day. During the summit Howard Locke and I shared profound insights into how cyber threats are evolving with the rise of Quantum technology, shedding light on the strategies adopted by malicious actors in this realm. For further exploration of this intriguing surveillance approach and its far-reaching implications, delve into the insightful article shared by Solomon J, featuring Gharun S. Lacy, Deputy Assistant Secretary in the Directorate of Cyber and Technology Security at the U.S. Department of State [Link to the article](https://lnkd.in/e47q9r5H)

10

2 Comments

There’s a big difference between drowning in data and actually using it to act. Follow the link to learn how SAIC is turning scattered information into clarity that helps operators make faster, better decisions — whether that’s in space, on the ground or in joint missions.

4

Thank you, Chris Kraft for sharing this! It is a great study for federal IT leaders to review. The CAISI findings highlight something we've been discussing with agencies: the architecture surrounding your AI matters as much as the models themselves. Key takeaways: On security: The 12x difference in agent hijacking susceptibility underscores why agencies need validation and orchestration layers that sit above any individual model. When processing sensitive data for investigations, eDiscovery, or breach response, you can't rely solely on model guardrails. On cost: The 35% cost differential for similar performance shows agencies that architect for model flexibility rather than vendor lock-in can optimize as the landscape evolves. On performance: The 20-80% variance on cyber tasks shows capabilities are moving targets. What works today may not tomorrow. The common thread: data quality and orchestration are your constants. When you curate unstructured data properly and build orchestration layers that let you leverage multiple models with consistent security controls, you're not betting on any single vendor's trajectory. For agencies dealing with sensitive data such as investigative work, compliance, and mission-critical analysis...having a platform that maintains data sovereignty while giving flexibility to bring the right AI capability to each use case isn't just nice to have. This study shows it's foundational. Thanks for sharing - important reading for AI governance in federal environments.

4