IT News Analysis

1

OpenAI, Anthropic, Google API Flaw Let Weaker AI Models Decode Stronger Models' Reasoning

A newly disclosed flaw in the way OpenAI, Anthropic, and Google carried hidden AI reasoning between API calls let researchers recover internal reasoning and secrets from session logs, including API keys and passwords. The weakness affected encrypted reasoning objects used by the providers' reasoning APIs, where a block created in one session could be replayed into another and, during testing, even handed to a weaker model in the same provider family to make it reveal the hidden content. The team behind the paper  Stealing Reasoning Traces from Proprietary LLM APIs  demonstrated four abuse paths: stealing proprietary reasoning for  model distillation , extracting private data from other users' published traces, recovering harmful content concealed behind a safe visible answer, and hiding prompt injections inside opaque reasoning blocks. Across 6,708 public agent trajectories, the team decoded 315,320 thinking blocks. After excluding benchmark sources, it counted 704...

Aug 12, 2026

2

Enterprise Defenses Recovered at the Edge and Collapsed Inside

Enterprise defenses are tuned to catch the attacks that make noise. This year's data shows attackers winning by making none. According to Picus Labs' new Blue Report 2026 , which measured more than 338 million real attack simulations across actual client production environments in the first half of 2026, defenses are having one of their strongest years yet. Average prevention effectiveness climbed from 62% to 69% , matching its 2024 peak, and logging reached a four-year high of 58% . The good news: The recovery is real. The bad news: It's taking place almost exclusively at the perimeter. The report's sharper finding is what happens after that perimeter is crossed . Inside, the picture inverts : defenses that look strong from the outside turn soft , and are the softest of all against the quiet moves, the reconnaissance and credential theft that precede every serious breach.  This is a fault line that runs through the entire report. A vulnerable interior behin...

Aug 12, 2026

3

AI Is Flooding Security with Bugs Nobody Proved

Stephen Sims on why unproven AI findings pile up as a triage burden. Read the full SANS blog.

4

Adobe Patches Three CVSS 10.0 ColdFusion and Campaign Classic Flaws

Adobe has shipped updates to address multiple critical security vulnerabilities impacting ColdFusion, Commerce, and Campaign Classic that, if successfully exploited, could result in arbitrary code execution and privilege escalation. The most severe of the flaws are listed below - CVE-2026-48362 (CVSS score: 10.0) - An operating system command injection vulnerability in ColdFusion that could lead to arbitrary code execution (Fixed in 2025.0.12 and 2023.0.23) CVE-2026-48273 (CVSS score: 9.9) - An eval injection vulnerability in ColdFusion that could lead to arbitrary code execution (Fixed in 2025.0.12 and 2023.0.23) CVE-2026-71384 (CVSS score: 9.6) - An incorrect authorization vulnerability in ColdFusion that could lead to an application denial-of-service (Fixed in 2025.0.12 and 2023.0.23) CVE-2026-71362 (CVSS score: 9.1) - An incorrect authorization vulnerability in Commerce that could lead to privilege escalation CVE-2026-71398 (CVSS score: 10.0) - An incorrec...

Aug 12, 2026

5

Attackers Exploit VMware vCenter Vulnerability to Gain Persistent Remote Access

Threat actors have begun to actively exploit a recently patched critical security flaw in Broadcom VMware vCenter, according to new findings from QUIRSO. The vulnerability in question is CVE-2026-59310 (CVSS score: 9.8), a directory-traversal vulnerability in the VMware vCenter server that a malicious actor with network access can exploit to execute arbitrary code. Patches for the flaw were released by Broadcom late last month. The German cybersecurity company said it discovered the activity following an incident response engagement. The attack chain is said to have exhibited path traversal activity consistent with the flaw, followed by the deployment of a malicious cron job to establish persistence on the host using reverse_ssh, an open-source tool used for setting up SSH connections to threat actor-controlled infrastructure. Compromised systems identified by QUIRSO were found to first establish contact with the attacker's domains on August 3, five days after Broadcom pub...

Aug 12, 2026

6

Malicious LiteLLM Releases Tied to Trivy Hack May Have Exposed 2,100+ Organizations

Two malicious LiteLLM releases sat on PyPI for about 40 minutes in March carrying credential-stealing code capable of harvesting cloud keys, SSH keys, Kubernetes tokens, database passwords, and other secrets from systems that installed them. Threat intelligence firm CloudSEK now says a dataset it obtained, built from roughly 434,000 files the attackers captured, maps potential exposure to more than 2,500 organizations. Those totals are not a victim count. CloudSEK told The Hacker News the material came from confidential intelligence sources and consists of captured loot and log files it assessed as belonging to the campaign, not data gathered from the organizations it names. The files were taken, in other words. CloudSEK has published the dataset as  a public lookup , searchable by name or domain and filterable by confidence. Each row gives an organization's name and domain, a count of secrets exposed, a count of runs, and a label reading High or Medium. What a high-co...

Aug 12, 2026

7

Exploit Time Just Dropped From 30 Days to 30 Minutes

Download the 11-step checklist CISOs use to close exposure windows first.

8

SAP Commerce Cloud Flaw Could Let Unauthenticated Attackers Execute Arbitrary Code

SAP has released patches to address a maximum-severity security flaw impacting Commerce Cloud (Data Hub Adapter) that could result in arbitrary code execution. The vulnerability, assigned the CVE identifier CVE-2026-58231 , is rated 10.0 on the CVSS scoring system. It has been described as a case of insufficient authorization checks and input validation. "SAP Commerce Cloud allows an unauthenticated attacker to abuse a default authentication client and submit specially crafted input to certain functions lacking sufficient validation," according to a description of the flaw on CVE.org. "Successful exploitation could enable arbitrary code execution and compromise internal components, resulting in high impact on confidentiality, integrity, and availability of the application." SAP security company Onapsis has urged customers to patch to a fixed Commerce Cloud release and then re-deploy the updated SAP Commerce Cloud version. As a temporary workaround until a...

Aug 12, 2026

9

ShieldBreak Zero-Day PoC Claims Microsoft Defender Patch Bypass With SYSTEM Access

The security researcher going by the name Chaotic Eclipse (aka INFINITE NIGHTMARE, MSNightmare, and Nightmare-Eclipse) has released a proof-of-concept (PoC) for a new Microsoft zero-day called ShieldBreak . The vulnerability, rooted in Microsoft Defender for Windows, demonstrates a patch bypass for CVE-2026-50656 (CVSS score: 7.8), otherwise known as RoguePlanet . RoguePlanet has been described as a race condition that, if successfully exploited, could grant an attacker the ability to spawn a shell with SYSTEM-level privileges, enabling them to run arbitrary code or perform unauthorized actions. Although it was first disclosed by the researcher in June 2026, a patch for the vulnerability was not released by Microsoft until almost a month later. The tech giant described it as a privilege escalation issue in the Microsoft Malware Protection Engine ("mpengine.dll"). Soon after, Chaotic Eclipse said the "defense-in-depth updates" introduced by Microsoft to ad...

Aug 12, 2026

10

Cisco ASA and FTD Flaw Exploited in the Wild Can Trigger Remote DoS

Cisco has warned that a new vulnerability impacting Secure Firewall Adaptive Security Appliance (ASA) Software and Secure Firewall Threat Defense (FTD) Software has been exploited in the wild. The high-severity flaw, tracked as CVE-2026-20349 (CVSS score: 8.6), is a case of insufficient error checking when processing HTTP requests that could allow an unauthenticated, remote attacker to trigger a denial-of-service (DoS) condition. "An attacker could exploit this vulnerability by sending a crafted HTTP request to the Remote Access SSL VPN service on an affected device," Cisco said in a Tuesday advisory. "A successful exploit could allow the attacker to cause the affected device to reload, resulting in a DoS condition." The security defects impact devices running a vulnerable version of Secure Firewall ASA Software or Cisco Secure FTD Software and have one or more of the vulnerable configurations listed below - IKEv2 Remote Access VPN (with client services...

Aug 12, 2026