Daily Edition · Page 5

THE DAILY INDEX

Technology, AI, software, security, crypto and business

ATTACKS & VULNERABILITIES

TLDR InfoSec

MISCONFIGURED LICENSE PLATE READERS ARE LEAKING DATA IN REAL TIME

Security Researcher Matt Brown purchased and reverse engineered the firmware of a Motorola automated license plate reader (ALPR) system. Brown used a unique error page to find 150 Internet exposed Motorola ALPR systems that did not implement any form of authentication. Wired journalists confirmed that these cameras are deployed by law enforcement and are leaking license plate number, car location, make and model, and other metadata from thousands of cars in just 20 minutes of sample time.

CROSS CACHE ATTACK CHEATSHEET

This article describes a technique called "cross-cache attack" in Linux kernel exploitation that allows transferring a Use-After-Free (UAF) vulnerability from one object to another, even if they're allocated from different slabs. The attack involves a five-step process: spraying object-A sharing a slab with object-V, allocating a victim object-V, freeing object-V, freeing specific objects to trigger slab recycling through the buddy system, and finally spraying object-B to overlap with the freed object-V's memory. The article provides detailed code examples and experimental results, including a timing-based technique called SLUBStick that helps determine the state of slab allocation in the kernel.

11 THOUGHTS ON “A DAY IN THE LIFE OF A PROLIFIC VOICE PHISHING CREW”

Brian Krebs has notes on a sophisticated voice phishing operation called "Crypto Chameleon" that is exploiting Apple and Google's legitimate services to scam cryptocurrency owners, using a coordinated team approach and phishing panels to steal millions from victims including billionaire Mark Cuban, while operating openly on messaging platforms like Discord and Telegram where the criminals frequently betray each other.

800+ “POTENTIAL VULNERABILITIES AND BIASES” MAY AFFLICT AI-INFUSED MEDICAL SERVICES

The Chief Digital and Artificial Intelligence Office (CDAO) led an effort in coordination with the Crowdsourced AI Red-Teaming Assurance Program and healthcare agencies to evaluate the usage of LLMs in military medical services. The over 200 participants evaluated three popular LLMs for a clinical note summarization and advisory Chatbot. The effort was concluded with over 800 potential vulnerabilities and biases identified with suggested improvements to reduce the number in future efforts.

TLDR IndexPage 5