TLDR InfoSec

TLDR INFORMATION SECURITY 2025-01-21

5 sections15 entrieslinks blockcomplete parse

Parse notes

  • Subscription/account footer truncated before parsefooter_truncated

01ATTACKS & VULNERABILITIES

HPE INVESTIGATES BREACH AS HACKER CLAIMS TO STEAL SOURCE CODE

A hacker going by the name IntelBroker claimed to have breached HPE. The hacker is selling an archive that allegedly contains certificates (private and public keys), Zerto and iLO source code, Docker builds, and old user personal information used for deliveries. HPE disabled relevant credentials but claimed that it had not discovered any evidence of a breach.

bleepingcomputer.com2 min read

DONOT TEAM LINKED TO NEW TANZEEM ANDROID MALWARE TARGETING INTELLIGENCE COLLECTION

A hacking group called DoNot Team has been linked to a new Android malware named Tanzeem that targets specific individuals for intelligence collection. The malware masquerades as a chat app but actually collects sensitive data and establishes connections to a command-and-control server. Cyfirma notes that the group uses push notifications to deploy additional malware and ensure persistence on targeted devices for ongoing intelligence gathering.

links.tldrnewsletter.com2 min read

TELEGRAM-BASED “SNEAKY 2FA” PHISHING KIT TARGETS MICROSOFT 365 ACCOUNTS

A new phishing kit called Sneaky 2FA targets Microsoft 365 accounts using evasion tactics and a Telegram platform. The kit is offered as a service on Telegram and incorporates techniques to avoid detection, posing a significant threat to cybersecurity. Organizations are advised to monitor threats closely and enhance their security measures to combat this growing risk.

hackread.com3 min read

02STRATEGIES & TACTICS

AI-SUPPORTED SPEAR PHISHING FOOLS MORE THAN 50% OF TARGETS

A group of researchers tested spear-phishing emails that were created by AI agents based on Claude 3.5 Sonnet and GPT-4o which crawled the Internet for information on the targets. The researchers achieved a click-through rate (CTR) of 54% with the AI agents compared to 12% in the control group. Human-crafted spear-phishing emails also achieved a 54% CTR but at 30x the cost of the AI-generated ones, whereas emails crafted with AI and human input achieved a CTR of 56% with 4x the cost.

malwarebytes.com3 min read

PREVENTING UNINTENDED ENCRYPTION OF AMAZON S3 OBJECTS

AWS detected unauthorized encryption activities in Amazon S3 due to malicious actors with valid credentials. Customers can prevent such incidents by using short-term credentials, implementing data recovery procedures, and monitoring access patterns. AWS notes that blocking of use of SSE-C encryption can also enhance security against unauthorized access.

aws.amazon.com6 min read

INVESTIGATING AN "EVIL" RJ45 DONGLE

Michał Zalewski reverse-engineered a suspicious RJ45 dongle from China and revealed it was not malicious, but rather a device using a software-defined CD-ROM feature to install drivers. The device contained a chip that emulated a CD-ROM to simplify driver installation, dispelling concerns of potential malware. The investigation highlighted the importance of research and patience in understanding complex hardware systems.

lcamtuf.substack.com5 min read

03LAUNCHES & TOOLS

IAM SIMULATE

An AWS IAM simulator that provides detailed explanations of how a policy was evaluated. It supports various IAM features like identity policies, resource policies, and more. The simulator automatically validates inputs and will soon include additional features like session policies and support for anonymous requests.

github.comGitHub repository

ROLES

Roles is a tool written in Golang for testing unauthenticated AWS IAM role enumeration.

github.comGitHub repository

GITXRAY

GitXRay is a security tool for GitHub repositories that uses public GitHub APIs for various purposes like OSINT and Forensics. It can detect sensitive information, identify threat actors, and find fake or infected repositories. Users can install and run GitXRay easily to enhance repository security and gather valuable insights.

github.comGitHub repository

04MISCELLANEOUS

MICROSOFT FIXES WINDOWS SERVER 2022 BUG BREAKING DEVICE BOOT

Microsoft has released a fix for a bug that was causing Windows Server 2022 systems with two or more NUMA nodes to fail at startup. NUMA nodes are commonly used to optimize CPU and memory usage. Microsoft also announced that it is working on a fix for a Windows Server 2025 issue that is causing iSCSI environments to fail to boot.

links.tldrnewsletter.com2 min read

VENOMOUS BEAR APT ADVERSARY SIMULATION

This blog discusses a simulation of an attack by the Venomous Bear APT group that targeted countries like the USA and Germany from at least 2020. The attackers used a fake Microsoft Windows Time service to install a backdoor that allowed them to exfiltrate files and communicate with a command and control server. The simulation includes stages for installing the backdoor using a .bat file, developing the backdoor DLL, and creating a listener script for receiving commands from the backdoor.

links.tldrnewsletter.com3 min read

HOW GOOGLE DOES IT: MAKING THREAT DETECTION HIGH-QUALITY, SCALABLE, AND MODERN

In this blog post, the Google security team describes their security approach, which involves automated threat detection with a focus on minimizing dwell time. Collaboration and building an asset inventory are key strategies to enhance threat detection capabilities. Security engineering is essential, and Google emphasizes writing code for effective threat response.

links.tldrnewsletter.com7 min read