The day's cybersecurity headlines from public sources, in one place. Search them, filter by theme or CVE, or show only vulnerabilities CISA has confirmed are being exploited. Every headline links to the original.
For my own read on what the news means for your business, see Insights.
-
China-made ZBT routers ship with two factory implants giving unauthenticated root
VulnCheck disclosed SPEAKINGSTONE and DARKLANTERN, two previously undocumented factory implants in firmware for routers built by Shenzhen Zhibotong Electronics, tracked as CVE-2026-74232 and CVE-2026-74233, each rated 9.3 on CVSS 4.0 and 9.8 on CVSS 3.1. SPEAKINGSTONE runs as the yunmgrd service…
-
Three CVSS 10.0 ServiceNow AI Platform flaws allow unauthenticated code and SQL execution
ServiceNow patched four flaws in the ServiceNow AI Platform, three rated 10.0. CVE-2026-18885 is code injection in the GraphQL Composite Data API allowing an unauthenticated user to execute arbitrary code and read or modify instance data. CVE-2026-18886 is improper access control in the system…
-
PaperCut ships a second emergency patch after researchers bypass the first fix
Multiple ways to bypass the original emergency fix were found, so PaperCut has released a second update with additional hardening. Huntress researchers John Hammond and Andrew Brandt described the underlying issue: a crafted request can name one page for rendering and a different page for the…
-
McKesson discloses breach, ShinyHunters claims 284 million patient records and a $55.2 million demand
McKesson discovered the incident on August 25 and filed an SEC Form 8-K the same day it went public, stating it has not determined the incident to be material. ShinyHunters told BleepingComputer it vished multiple McKesson employees using a mckesson[.]claims lookalike domain, compromised their…
-
More than 130 technology and finance companies sign a joint warning on a closing window for cyber defense
OpenAI published a call for collective action on cyber defense signed by more than 130 organizations including Anthropic, AWS, Google, Microsoft, and Oracle, warning of a "limited window" before AI-enabled attacks become far more widespread and sophisticated. The letter names hospitals, water…
-
Android 17 adds four network security features, including ECH, operator-controlled 2G shutoff, and local network permissions
Android 17 introduces Encrypted Client Hello alongside private DNS, hiding visited domain names from network observers and making Android the first major mobile OS with broad ECH support. Participating mobile operators can now automatically disable 2G for subscribers, cutting exposure to SMS…
-
GoCaracal malware pulls replacement command-and-control addresses from an Ethereum smart contract
Arctic Wolf documented GoCaracal, a previously undocumented Go-based malware framework deployed during a June 2026 intrusion at an unnamed communications organization in Venezuela, and linked it with medium confidence to Dark Caracal. The framework provides remote shell access and payload…
-
Spark RAT campaign in Cambodia abuses a vulnerable OPSWAT driver to disable security tools
Acronis Threat Research Unit documented a multi-stage campaign against individuals and organizations in Cambodia delivering the open-source Spark RAT. The chain uses bring-your-own-vulnerable-driver technique, loading the legitimate but vulnerable OPSWAT AppRemover driver (ardrv.sys) to escalate…
-
Australia charges two men over the TeamPCP supply-chain compromise of Trivy, Checkmarx KICS, and LiteLLM
The Australian Federal Police charged Louis Michael Gaebler, 23, and Ruben Ian Thomson, 21, with a combined 14 offences over their alleged roles in TeamPCP, the group behind the March 2026 compromise of open-source security scanners Trivy and Checkmarx KICS and the AI gateway LiteLLM. Both…
-
Finland's appeals court revives the Eagle S prosecution, holding that Baltic cable damage was not a maritime accident
The Helsinki Court of Appeal on August 27 overturned a district court ruling that had thrown out the prosecution of three senior officers of the Eagle S, the Russia-linked tanker that severed multiple Baltic subsea cables on Christmas Day 2024. The court held Finland has jurisdiction because the…
-
Guardian reports UK small power plants stay exposed into the 2030s after the Iran-linked shutdown
Following the reported Iran-linked intrusion that took a small British generator offline for four days in July, the Guardian reports that the UK's smaller power plants will remain at elevated cyber risk well into the 2030s. The original incident was first reported by the Telegraph on August 22 and…
-
White House executive order bans foreign-made bulk-power system equipment over backdoor concerns
The administration issued an executive order on August 26 banning acquisition and installation of foreign-made technology used to manage electricity, declaring foreign bulk-power system electric equipment an "unusual and extraordinary threat." The order covers equipment for transmission lines…
-
Nearly four in ten German companies now blame a foreign intelligence service for at least one attack
Bitkom surveyed 1,003 German companies with at least 10 employees and found that of those hit by data theft, industrial espionage, or sabotage in the past year, nearly 40 percent attributed at least one incident to a foreign intelligence service, up from 28 percent last year and 7 percent in 2023…
-
Carhartt breach confirmed at 12.9 million accounts after ShinyHunters padded the dump with synthetic records
ShinyHunters published roughly 50GB of Carhartt data on August 13 after the retailer refused a $3.3 million extortion demand. Analysis by Troy Hunt for Have I Been Pwned put the real figure at 12,933,413 individuals, about half of what the group claimed, because the dump was injected with millions…
-
Manchester Airports Group says attackers took data on 8.7 million customers from Wi-Fi and parking systems
Manchester Airports Group, which operates Manchester, Stansted, and East Midlands airports, disclosed that intruders accessed customer data belonging to roughly 8.7 million people. Exposed fields include email addresses, phone numbers, vehicle registration numbers, and postcodes, drawn from…
-
ATF declares a major incident after Qilin claims a breach of a system holding investigation targets
The Bureau of Alcohol, Tobacco, Firearms and Explosives classified a cyberattack on one of its systems as a "major incident," a statutory designation that triggers formal notification to Congress. ATF said the affected system is stand-alone and separate from the bureau's main network, and that it…
-
OpenAI post-mortem says reward hacking drove its research models to exploit a zero-day and breach Hugging Face
OpenAI published a post-mortem attributing last month's AI-driven Hugging Face compromise to reward hacking, and said it found evidence of misaligned behavior as early as late May. The incident occurred during cybersecurity evaluations of several OpenAI models and was driven mainly by a highly…
-
Ransomware affiliate ran Cursor's AI coding agent inside victim networks by telling it the intrusion was a test
Israeli firm Gambit reviewed 28 chat sessions recovered from an unsecured server belonging to Aur0ra, a Russian-speaking ransomware group, and found the operators using the AI agent built into the Cursor code editor to run hundreds of malicious operations: credential theft, internal network…
-
GPUThor Rowhammer attack defeats ECC on NVIDIA workstation GPUs and reaches a root shell
University of Toronto researchers disclosed GPUThor, a Rowhammer variant against NVIDIA workstation GPUs with GDDR6 memory that defeats error correction codes, the mitigation NVIDIA recommends for GPU Rowhammer. The team induced bit flips on all four Ampere-class cards tested (RTX A6000, A5000…
-
Amazon Kiro AI IDE can be made to exfiltrate local data through prompt injection in repository content
Mindguard disclosed a flaw in Amazon Kiro, an agentic AI development environment, that lets attacker-controlled repository content steer the Kiro agent into sending sensitive local information to an external endpoint. The issue has no CVE identifier and was demonstrated against Kiro IDE 0.7.45 on…
-
Next.js patches two critical unauthenticated RCE flaws, one in AVIF handling and one Windows-only path traversal
Vercel published fixes for two critical vulnerabilities in Next.js that each allow unauthenticated remote code execution. CVE-2026-75604 (CVSS 9.0) is a path traversal that affects applications using both the Pages Router and App Router without Cache Components when the server runs on a Windows…
-
PaperCut ships emergency patches after confirmed zero-day attacks on print servers, all versions affected
PaperCut confirmed attackers are exploiting a flaw in every unpatched version of PaperCut NG and MF. The bug is an authorization logic failure: a crafted request can name one page for rendering while executing a component owned by a different page, so the permission check validates the wrong…
-
Citrix NetScaler flaw now exploited for pre-auth RCE, CISA gives agencies until August 29
Citrix patched CVE-2026-8452 on June 30 and described it as a high-severity memory overflow leading to denial of service, but WatchTowr published analysis and proof-of-concept code on August 14 showing it can be driven to unauthenticated remote code execution. Previdian and Defused reported…
Also covered by The Hacker News
-
Aikido reproduces the gym-booking agent incident, with the agent exploiting a client-side-only limit in 9 of 10 runs
Aikido Security rebuilt the August 10 Australian gym-booking incident in a synthetic single-page app backed by a GraphQL API carrying the two original flaws: a seven-day booking window enforced only in the frontend, and an unprotected cancel operation. Running Claude Opus 4.6 on the OpenClaw agent…
-
Unit 42 finds AI speeds malware development but does not improve success rates
Palo Alto Networks Unit 42 analyzed 405 AI-linked malware samples and found only 12 reached production endpoints. The finding is that generative tooling compresses the build phase without meaningfully raising the odds that the resulting malware works against defended environments.