Security Round-Up

CONTENTS

TLDR InfoSec
Hacker News ( 16 )
Check Point Research
Cisco Talos ( 2 )
Bleeping Computer ( 11 )
CISA
MS-ISAC CYBERSECURITY ADVISORY
Advisories ( 2 )
DataBreaches.net ( 10 )
CVEMon Intruder ( 10 )
Embrace The Red
Graham Cluley
Publications | Hacking Lab ( 2 )
Schneier on Security ( 3 )
Securelist
Talos – Vulnerability Reports
Troy Hunt ( 2 )
Veeam ( 4 )

---

TLDR InfoSec

08/02 TOC Ruby On Rails ActiveStorage Flaw , Amgen Cloud Breach , HuggingFace Se…

---

Hacker News

08/04 TOC DOUBLECUP Uses ClickFix and Cached PNGs to Deliver CountLoader and Dev…

A new Russian loader-as-a-service (LaaS) codenamed DOUBLECUP has been using ClickFix lures as a way to stage malware-laced PNG images in victims’ browser cache and ultimately deliver CountLoader and a previously undocumented remote access trojan called DeviceManager. “The first stage drops a steganographic PNG image into the browser’s cache, retrieves its hidden content, and executes the second

08/04 TOC CISA Adds Exploited N-able N-central Flaw to KEV After Customer Compro…

The U.S. Cybersecurity and Infrastructure Security Agency (CISA) on Monday added a high-severity security flaw impacting N-able N-central to its Known Exploited Vulnerabilities (KEV) catalog following reports of active exploitation in the wild. The vulnerability, tracked as CVE-2026-18577 (CVSS score: 8.2), is a case of incomplete patching for CVE-2026-18556 (CVSS score: 8.2) that allows

08/03 TOC 18 Malicious npm Packages Deliver Cross-Platform RAT to Alibaba Tool U…

Cybersecurity researchers have discovered a new set of malicious npm packages that target users of Alibaba developer tools with a cross-platform remote access trojan (RAT) as part of a sophisticated, targeted software supply chain attack targeting Chinese-speaking environments. One of the packages in question is “lib-mtop,” an unscoped package with the same name as a private Alibaba package

08/03 TOC Google Password Manager Attacks Could Let Malware Hijack Passkey-Prote…

Malware running as an ordinary user on a Windows machine can sign into a victim’s passkey-protected accounts without a fingerprint, a PIN, or anything at all appearing on the victim’s screen. Unit 42 detailed three attack paths against Chrome’s Google Password Manager cloud authenticator, which it calls Pass-ta-key, Silver Pass-ta-key and Golden Pass-ta-key; the strongest targets the master key

08/03 TOC INC Ransomware Emerges as Dominant Actor Exploiting SonicWall SMA 1000…

The INC Ransomware operation has emerged as the “dominant threat actor” exploiting the recently disclosed security flaws in SonicWall Secure Mobile Access (SMA) 1000 series VPN appliances. In a report published over the weekend, Resecurity said it observed the INC Ransomware accelerating its activity since the beginning of August 2026, listing multiple victims on its data leak site. Per

08/03 TOC Weekly Recap: Rogue AI Models, $88M Bitcoin Theft, Water-System Attack…

This week kept coming back to permission. A model crossed a boundary. A wallet trusted bad randomness. Webmail kept an intruder around. Public systems, package feeds, hotel networks, and login flows all gave away more than intended. Some of it was clever. Most of it was just access left lying around: old bugs, exposed gear, poisoned dependencies, weak defaults, and tooling that moved from

08/03 TOC FOMO in the SOC: Where AI Platforms like Claude Actually Fit

AI is moving incredibly fast, and every security leader is feeling the pressure to keep up. AI platforms like Claude, Codex and Cursor are already helping security teams write detections, investigate alerts, summarize incidents, and automate repetitive work. The conversation has evolved from whether AI belongs in the SOC, to where each type of AI delivers the most value. With so many new AI

08/03 TOC Chinese Threat Actor Uses Leaked DarkSword Kit to Deploy GHOSTBLADE on…

An unknown Chinese-speaking threat actor has been observed running a campaign targeting Apple iOS devices by leveraging a publicly leaked version of the DarkSword exploit kit. Attack surface management platform Censys said it identified the threat actor running more than 100 web properties, most of which are fake Amazon Web Services (AWS) sign-in pages on a domain that also hosts the exploit

08/03 TOC PNLD Breach Exposes U.K. Police and Government Contact Details on Dark…

The Police National Legal Database (PNLD) has confirmed that police, government and customer contact information was compromised and published on the dark web. The data included names, organisations and work email addresses belonging to police officers, police staff, criminal justice professionals, government partners and customers. The incident, identified on July 26, also exposed some names

08/03 TOC Thermo Fisher Patches Flaw That Could Make DNA File Tampering Nearly U…

Thermo Fisher Scientific has patched a flaw in select Applied Biosystems human identification software that could allow data files to be altered before analysis software loads them. The vendor’s July 31 security bulletin says nearly undetectable changes to .fsa and .hid outputs could occur if laboratory controls are circumvented. Thermo Fisher tracks the issue as CVE-2026-17583 and rates it

08/03 TOC N-able Says Attackers Take Over N-central Servers After Initial Fix Pr…

N-able said attackers exploited an authentication bypass in N-central to gain remote administrative access and reach the customer systems managed through those servers. Its first fix was incomplete. CVE-2026-18577 affects N-central builds prior to 2026.3.1.7. N-able shipped build 2026.3.1.7 on August 2 as the first unaffected version. N-central is the remote monitoring and management platform

08/03 TOC Hugging Face Diffusers Flaws Could Let Model Repositories Execute Arbi…

Three high-severity security flaws have been disclosed in Hugging Face’s Diffusers library that could allow crafted model repositories to stealthily execute arbitrary code on machines that load it, opening the artificial intelligence (AI) supply chain to security risk. “These vulnerabilities are bypassing trust_remote_code, the safeguard designed to stop unreviewed code from running in the

08/01 TOC Coldcard Hardware Wallet Flaw Linked to $70 Million Bitcoin Theft in 4…

An attacker drained 1,196 Bitcoin addresses in 41 minutes on July 30, taking 1,082.65 BTC worth about $70.2 million at the time. Galaxy Research mapped the sweep and tied it to a firmware flaw in Coldcard, the Bitcoin-only hardware wallet made by Canadian firm Coinkite. A March 2021 firmware integration error routed seed generation to a deterministic software pseudorandom number generator (PRNG

08/01 TOC Hackers Poison Adform Script to Swap Crypto Wallet Addresses Across Cu…

Attackers modified a JavaScript file served by advertising technology company Adform, turning it into a browser-side tool that rewrites cryptocurrency wallet addresses. Adform detected the incident on July 27, 2026, removed the malicious code, notified affected clients, and reported it to authorities. Anyone who visited a site carrying the affected script on July 27 and copied a Bitcoin,

08/01 TOC Adobe Campaign Classic CVSS 10.0 Flaw Could Run Code Without User Inte…

Adobe has released security updates to address a maximum-severity security flaw in Campaign Classic (ACC), its enterprise-focused marketing automation platform, that could result in arbitrary code execution. The vulnerability, tracked as CVE-2026-48449, carries a severity score of 10.0 on the CVSS scoring system. It has been described as a case of incorrect authorization that could result in

08/01 TOC Hijacked Hotel Wi-Fi Pushes Fake Updates to Deliver Surveillance Malwa…

A fake browser update served over hijacked hotel Wi-Fi has been used to deliver CornFlake, a remote access trojan (RAT) that can capture webcam images, microphone audio, and keystrokes, Microsoft said in its latest report. Researchers track the operation as CaptiveCrunch and attribute it to Storm-2945. It assesses Storm-2945 to be an operational sub-cluster of Midnight Blizzard, also known as

---

Check Point Research

08/03 TOC 3rd August Threat Intelligence Report

For the latest discoveries in cyber research for the week of 27th July, please download our Threat Intelligence Bulletin. TOP ATTACKS AND BREACHES Minnesota IT Services has confirmed coordinated cyberattacks affecting more than 30 community water utilities across the state. The incidents briefly disrupted a treatment plant in Braham and affected industrial control systems. Officials reported […]

The post 3rd August – Threat Intelligence Report appeared first on Check Point Research.

---

Cisco Talos

08/04 TOC Keep going, bro. Youve got this! A data-driven look at how adversaries…

Talos has collected prompt logs from threat actor endpoints running various applications, such as Claude Code, CodeX, Cursor, or Gemini. This blog is an analysis of the ways we’ve seen bad actors leveraging cloud-based AI.

08/03 TOC [Webinar] Tales from the Frontlines: An exclusive briefing on Q2 incid…

Register for an exclusive, unrecorded 30-minute webinar to review the most high-impact incidents Talos IR faced in Q2.

---

Bleeping Computer

08/03 TOC Hotel Wi-Fi attacks use custom malware to breach Microsoft 365 account…

Microsoft has linked a global campaign targeting hospitality Wi-Fi networks to the Russian threat actor Midnight Blizzard, also known as APT29. […]

08/03 TOC New Pass-ta-key attacks let malware hijack Google-synced passkeys

Security researchers have discovered three attacks that allow malware on already-compromised Windows devices to abuse Google Password Manager’s synced passkeys to take over accounts, bypass user verification, and extract passkey private keys. […]

08/03 TOC New DOUBLECUP ClickFix service hides malware in browser cache images

A new Russian loader-as-a-service named DOUBLECUP uses ClickFix attacks to hide malicious code in PNG images cached by victims’ browsers, ultimately delivering CountLoader to Windows and macOS devices and a new remote access trojan named DeviceManager to Windows systems. […]

08/03 TOC Fake Roblox Xeno script launcher pushes infostealer, RAT malware

Fake Xeno Executor installers are infecting unsuspecting Roblox players with malware that provides remote access and steals sensitive information. […]

08/03 TOC N-able warns of N-central auth bypass flaw exploited in attacks

N-able is warning customers that hackers are exploiting an authentication bypass vulnerability (CVE-2026-18577) affecting both hosted and on-premises N-central servers. […]

08/03 TOC ExfilSquad hackers leak info of over 100,000 UK police officers, staff…

A cyberattack on the U.K.’s Police National Legal Database (PNLD) has compromised contact data of more than 100,000 police officers and other criminal justice professionals. […]

08/03 TOC Inside the Underground Business of the Android BTMOB RAT malware

Flare researchers analyzed thousands of underground posts to examine how the BTMOB Android malware operation evolved into a fragmented ecosystem of resellers, source-code vendors, custom versions, and competing sales channels. […]

08/02 TOC OpenAI teases Astra, its next major AI model, after it solves 10 long-…

OpenAI has revealed Astra, an unreleased model designed to tackle complex, long-running tasks, after an internal version produced ten significant advances in mathematics and theoretical computer science. […]

08/02 TOC COLDCARD wallet RNG flaw likely linked to $88 million Bitcoin theft

A vulnerability in COLDCARD hardware wallet firmware allowed attackers to steal an estimated $88.6 million in Bitcoin from thousands of wallets whose seeds were generated using a flawed random number generator. […]

08/02 TOC Google Chrome may soon block New Tab hijacker extensions by default

Google is preparing a new Chrome security feature that would block policy-installed extensions from hijacking the New Tab page or changing the default search engine. […]

08/01 TOC Rails patches critical Active Storage flaw with RCE potential

A critical vulnerability in the Active Storage framework can allow an unauthenticated attacker to read arbitrary files from a Rails application, and potentially escalate to remote code execution (RCE). […]

---

CISA

08/03 TOC CISA Adds One Known Exploited Vulnerability to Catalog

CISA has added one new vulnerability to its Known Exploited Vulnerabilities (KEV) Catalog, based on evidence of active exploitation.  

• CVE-2026-18577 N-able N-central Authentication Bypass Using an Alternate Path or Channel Vulnerability

This type of vulnerability is a frequent attack vector for malicious cyber actors and poses significant risks to the federal enterprise.

Binding Operational Directive (BOD) 26-04: Prioritizing Security Updates Based on Risk establishes vulnerability management requirements for Federal Civilian Executive Branch (FCEB) agencies. BOD 26-04 reinforces the importance of the KEV Catalog and requires federal agencies to prioritize rapid remediation of high-risk vulnerabilities, specifically those identified by Common Vulnerabilities and Exposures (CVEs) listed in CISA’s KEV Catalog on publicly exposed assets that grant total control of the asset post-exploitation, while deferring action for lower-risk vulnerabilities. BOD 26-04 further establishes basic expectations for when agencies must check whether threat actors compromised the system before the patch was applied.

While BOD 26-04 applies only to FCEB agencies, CISA encourages all organizations to adopt risk-based vulnerability management and prioritize remediation of KEV Catalog vulnerabilities. CISA will continue to add vulnerabilities to the catalog that meet the specified criteria.

Aware of an exploited vulnerability not currently listed in the KEV Catalog? Submit it for potential addition through CISA’s KEV Nomination Form. Potential KEV additions must have a CVE ID, evidence of exploitation, and clear mitigation guidance.

---

MS-ISAC CYBERSECURITY ADVISORY

08/03 TOC Multiple Vulnerabilities in SolarWinds Web Help Desk Could Allow for A…

Multiple vulnerabilities have been discovered in SolarWinds Web Help Desk, the most severe of which could allow for authentication bypass. SolarWinds Web Help Desk software grants access to SolarWinds IT support, asset management, and knowledge base operations. A vulnerability in the Web Help Desk could allow an unauthenticated, remote attacker to bypass authentication and gain access. This does require the SAML 2.0 authentication method to be enabled. 

---

Advisories

08/03 TOC Cisco Secure Firewall Management Center Software Authentication Bypass…

A vulnerability in the web interface of Cisco Secure Firewall Management Center (FMC) Software could allow an unauthenticated, remote attacker to bypass authentication and execute script files on an affected device to obtain root access to the underlying operating system. 

This vulnerability is due to an improper system process that is created at boot time. An attacker could exploit this vulnerability by sending crafted HTTP requests to an affected device. A successful exploit could allow the attacker to execute a variety of scripts and commands that allow root access to the device. 

Note: If the FMC management interface does not have public internet access, the attack surface that is associated with this vulnerability is reduced.

Cisco has released software updates that address this vulnerability. There are no workarounds that address this vulnerability.

This advisory is available at the following link:
https://sec.cloudapps.cisco.com/security/center/content/CiscoSecurityAdvisory/cisco-sa-onprem-fmc-authbypass-5JPp45V2

This advisory is part of the March 2026 release of the Cisco Secure Firewall ASA, Secure FMC, and Secure FTD Software Security Advisory Bundled Publication. For a complete list of the advisories and links to them, see Cisco Event Response: March 2026 Semiannual Cisco Secure Firewall ASA, Secure FMC, and Secure FTD Software Security Advisory Bundled Publication.

Security Impact Rating: Critical
CVE: CVE-2026-20079

08/03 TOC Cisco Secure Firewall Management Center Software Static Credential Vul…

A vulnerability in the web interface of Cisco Secure Firewall Management Center (FMC) Software could allow an unauthenticated, remote attacker to log in to an affected device using a low-privileged account to access sensitive data within the impacted systems.

This vulnerability is due to the presence of static user credentials for a low-privileged account. An attacker could exploit this vulnerability by using the account to log in to an affected system. A successful exploit could allow the attacker to log in to the affected system and access sensitive data as the low-privileged user. 

Note: If the FMC management interface does not have public internet access, the attack surface that is associated with this vulnerability is reduced.  

Cisco has assigned this security advisory a Security Impact Rating (SIR) of High rather than Medium as the score indicates. The reason is that this vulnerability can be used with other Cisco Secure FMC Software vulnerabilities to elevate privileges.

Cisco has released software updates that address this vulnerability. There are no workarounds that address this vulnerability. 

This advisory is available at the following link:
https://sec.cloudapps.cisco.com/security/center/content/CiscoSecurityAdvisory/cisco-sa-fmc-static-cred-BET3Cjh

Security Impact Rating: High
CVE: CVE-2026-20316

---

DataBreaches.net

08/03 TOC KR: Seoul lawmaker criticizes 5,000-won compensation for 4.62 million-…

The Herald Business reports: Seoul Facilities Corp. has drawn criticism over its plan to offer 5,000 won [$3.50 USD] per affected user in response to a personal data breach involving about 4.62 million people, with questions mounting over whether the compensation is adequate. Seoul Metropolitan Council member Im Gyu-ho of the Democratic Party of Korea…

Source

08/03 TOC UK: Details of 100,000 police staff leaked on the dark web after hack

Bill Curtis reports: The full names and contact details for more than 100,000 police officers and staff have been leaked on the dark web after a hack, The Times can reveal. As part of a major security breach, hackers compromised data belonging to the Ministry of Defence (MoD), the Home Office, National Crime Agency (NCA),…

Source

08/03 TOC Cyberattack hits Liechtenstein, with 31,000 records stolen

DPA reports: The tiny principality of Liechtenstein has fallen victim to a major cyberattack in which the data of 31,000 people were stolen, the government said on Sunday. The country, which lies between Switzerland and Austria, has a population of around 41,000. The government said it had convened a crisis team led by Prime Minister…

Source

08/02 TOC A No-Logs VPN That Kept 58 Million Connection Logs: Inside the NotVPN …

They advertised and pinky swore “no logs.” But according to research by MysteriumVPN, they logged. Key takeaways from MysteriumVPN: A threat actor on the Altenen cybercrime forum is distributing a 17 GB SQL database they claim was stolen from SplitVPN (formerly NotVPN), a Russian VPN used to bypass internet blocks. The Mysterium research team obtained…

Source

08/02 TOC Brinks Home Confirms Data Breach Following ShinyHunters Claim

Guru Baran reports: Brinks Home, one of North America’s largest residential security providers, has confirmed that hackers breached its IT systems after the notorious ShinyHunters extortion group claimed responsibility for stealing nearly five million records tied to the company’s Salesforce environment. The confirmation comes after the threat actors listed “BH Security, LLC (brinkshome.com)” on their…

Source

08/02 TOC TN: Sumner County Schools provides limited update on data breach

Abbey Nutter reports: Sumner County Schools is still working through a reported network breach that forced the district to delay the start of the 2026-27 school year, officials told Main Street Media. The district reported the data breach during a meeting of the Sumner County Board of Education on July 21, one day after the…

Source

08/01 TOC Sixth Circuit to Rehear Case on FCC Data Breach Rules Case

Jake Neenan reports: A full panel of federal judges will rehear a case that upheld expanded telecom data breach rules. The Federal Communications Commission, now under Republican control, has indicated it’s likely to reverse the rules anyway. But industry groups and GOP lawmakers want the case’s precedent gone too. They told judges on the U.S….

Source

08/01 TOC The double extortion of a Russian ransomware threatens the medical rec…

Miguel Gomez reports: The biopharmaceutical company Diater, founded in Madrid in 1999, has appeared on the list of victims that the ransomware group DeadLock is disseminating on the dark web. The intrusion affects a company that manages particularly sensitive information of patients and healthcare professionals. The contrast lies in the type of data compromised and…

Source

08/01 TOC CareCloud Data Breach Impacts Over 350,000

Ionut Arghire reports: Healthcare information technology company CareCloud is notifying at least 350,000 people that their information was stolen in a data breach. The incident involved an electronic health record environment within the CareCloud Health division, which was disrupted on March 16, 2026. CareCloud’s investigation determined that hackers accessed one of its AWS environments between…

Source

08/01 TOC Suspected cyberattack disrupts Oceanside, California, school district …

DysruptionHub reports: A suspected cyberattack disrupted work email, internet access, Google Drive and other applications at Oceanside Unified School District in California as officials investigated and worked to restore service. The district confirmed a computer network disruption but did not identify its cause. NC Pipeline reported that a separate district text described the incident as a cyberattack….

Source

---

CVEMon Intruder

08/04 TOC CVE-2026-15929

Currently trending CVE – Hype Score: 64 – Improper neutralization of special elements used in an SQL command (‘SQL injection’) vulnerability in LG Electronics SmartShare allows SQL Injection. This issue affects SmartShare: through 2.3.1712.1202, which is supported on Microsoft Windows 10 and earlier versions.

08/04 TOC CVE-2026-58374

Currently trending CVE – Hype Score: 7 – In hostapd before 2.12, a missing bounds check in AP-mode Wi-Fi 7 (IEEE 802.11be) Multi-Link Operation (MLO) association request processing allows an unauthenticated attacker within wireless range to send a crafted management frame containing a malformed Multi-Link Element or …

08/04 TOC CVE-2026-61511

Currently trending CVE – Hype Score: 5 – vBulletin 5.x through 5.7.5 and 6.x through 6.2.1 contains an eval injection vulnerability in the vB5_Template_Runtime::runMaths() method within the template runtime that allows unauthenticated remote attackers to execute arbitrary PHP code by supplying crafted input through the …

08/04 TOC CVE-2026-16347

Currently trending CVE – Hype Score: 5 – MikroTik RouterOS contains a weakness in its API authentication handling that lacks effective safeguards against excessive login attempts. The system does not enforce meaningful rate-limiting, account lockout, or source-based restrictions, allowing repeated authentication …

08/04 TOC CVE-2026-32202

Currently trending CVE – Hype Score: 5 – Protection mechanism failure in Windows Shell allows an unauthorized attacker to perform spoofing over a network.

08/04 TOC CVE-2026-30623

Currently trending CVE – Hype Score: 4 – LiteLLM 1.18.10 contains a remote code execution vulnerability in its MCP server creation functionality. The application allows users to add MCP servers via a JSON configuration specifying arbitrary command and args values. LiteLLM executes these values on the host without …

08/04 TOC CVE-2025-71399

Currently trending CVE – Hype Score: 4 – Better Auth relies on better-call, which uses the rou3 router library. In affected versions of rou3, paths are normalized by removing empty segments, so /path, //path, and ///path resolve to the same route. In Better Auth versions prior to 1.4.5 (which bundles the fixed rou3), …

08/04 TOC CVE-2025-71400

Currently trending CVE – Hype Score: 3 – better-auth passkey versions before 1.4.0 contain an insecure direct object reference vulnerability in the passkey deletion endpoint that allows authenticated users to delete arbitrary passkeys by ID. Attackers with valid sessions can submit crafted requests to the …

08/04 TOC CVE-2025-15673

Currently trending CVE – Hype Score: 3 – The Import and export users and customers WordPress plugin before 2.4.3 does not restrict the path of a file it reads and displays during a CSV import, allowing high-privileged users to read arbitrary files on the server.

08/04 TOC CVE-2026-48449

Currently trending CVE – Hype Score: 1 – Adobe Campaign Classic (ACC) is affected by an Incorrect Authorization vulnerability that could result in arbitrary code execution in the context of the current user. Exploitation of this issue does not require user interaction. Scope is changed.

---

Embrace The Red

08/03 TOC LLM Heist: Hijacking LiteLLM for Traffic Interception, Key Theft, and …

LiteLLM is a popular AI gateway. It provides a unified interface to LLMs and simplifies governance. It also has access to the backend LLM provider keys.

All of that makes it a high-value target. Not only for IP and data theft, but also for response modification and tool invocation.

This post walks through a set of TTPs that red teams can integrate into authorized operations to demonstrate rerouting, interception, and modification of LLM traffic. We also cover things defenders can look out for.

---

Graham Cluley

08/04 TOC Fake IRS letters target cryptocurrency holders

Do you hold cryptocurrency? Have you received a letter telling you that you must register with a so-called “Digital Asset Compliance Portal”? If so, it’s time to hit the brakes, because it sounds like someone is trying to scam you. Read more in my article on the Hot for Security blog.

---

Publications | Hacking Lab

09/30 TOC Prism: A Multi-Team Orchestration of LLM Agents for Automatic Program …
08/31 TOC QueryHouse: Cross-DBMS Differential Testing with LLM and Query Transpi…

---

Schneier on Security

08/04 TOC Some Claude Chats Are Searchable on Google

And it’s personal information (alternate link):

The exposed data includes an AI-powered therapy app that someone appears to have vibe-coded, notes on meetings, and a dashboard someone made apparently to analyze medical billing data. Exposed chats reportedly include private cryptocurrency wallet keys and personal information like peoples’ addresses.

What seems to be the issue is a user setting about data sharing. Anthropic’s position is that it’s not their problem:

“We give people control over sharing their Claude conversations publicly, and in keeping with our privacy principles, we do not share chat directories or sitemaps with search engines like Google,” the company said in a statement. “These shareable links are not guessable or discoverable unless people choose to share them themselves. When someone shares a conversation, they are making that content publicly accessible, and like other public web content, it may be archived by third-party services.”…

08/03 TOC More on the OpenAI Agents Attack on Hugging Face

Hugging Face has published a detailed timeline of the attack. From the summary:

The agent was running an internal OpenAI cyber-capability evaluation based on the ExploitGym benchmark, which tasks an AI agent with finding and exploiting software vulnerabilities. OpenAI ran this on its own infrastructure, and the ExploitGym maintainers and their infrastructure had no involvement in the deployment or operation of that evaluation environment. As far as we were able to infer, across the course of being evaluated on this benchmark, the agent inferred that Hugging Face may host that benchmark’s models, datasets, and reference solutions. We believe the entire intrusion was, from the agent’s point of view, an attempt to cheat the evaluation: reach our production systems and steal the test solutions rather than solve the challenge on its own…

08/03 TOC The OpenAI Hack Shows the Genie Is Out of the Bottle

This essay originally appeared in Foreign Policy.

Earlier this month, two of OpenAI’s models broke out of their containment sandbox and attacked another AI company. The story is kind of wild. OpenAI was running security tests on two of its models: GPT-5.6 Sol and an unreleased model that is almost certainly GPT-6. In particular, it was running the ExploitGym benchmark, which measures how good a model is at turning security vulnerabilities into working exploits: basically, offensive cyberattacks.

Since these were internal tests, OpenAI locked those models in a secure sandbox that denied them access to the internet. But it was running the models without any safety filters that would prevent them from offensive cyber-actions. That meant that there was nothing to prevent the models from trying to …

---

Securelist

08/03 TOC An analysis of incidents at Brazilian educational institutions

Kaspersky expert provides statistics and details on several incident response cases at educational institutions in Brazil, as well as tips for schools and universities on how to stay safe.

---

Talos – Vulnerability Reports

08/03 TOC Microsoft Windows Cloud Files Mini Filter Driver CldiStreamCompleteReq…

---

Troy Hunt

08/03 TOC Welcoming the Nepalese Government to Have I Been Pwned

Today, we welcome the 47th government onboarded to Have I Been Pwned’s free gov service: Nepal. Their National Cyber Security Centre now has access to monitor Nepalese government domains against the data in HIBP. This gives the NCSC the ability to identify exposure across government email addresses and

08/02 TOC Weekly Update 515

Apparently, Aussies are so obsessed with coffee that it’s referred to as the coffee capital of the world down here (some bits, at least). “But what about Italy?” people ask. Having spent a lot of time in a lot of Italy, no, it’s just

---

Veeam

08/02 TOC Support Statement Exchange Web Services (EWS) Deprecation

All product features that currently utilize EWS will be transitioned to Microsoft Graph API, ensuring continued functionality.

08/02 TOC Granular sudo Permissions for Management of Veeam Agent for IBM AIX

Granular sudo Permissions for Management of Veeam Agent for IBM AIX

08/02 TOC Build Numbers and Versions of Veeam Backup for Microsoft Azure

This KB article lists all versions of Veeam Backup for Microsoft Azure and their respective build numbers.

08/02 TOC Build Numbers and Versions of Veeam Backup for AWS

Build Numbers and Versions of Veeam Backup for AWS