Hackers Used Claude to Hunt for Secrets in 1.8 Million Android Apps

, ,

Anthropic's latest threat intelligence report reveals that hackers used its AI model Claude to automate large-scale cyber operations, including harvesting credentials from 1.8 million Android apps and developing malware and phishing tools. The report highlights that AI is becoming an operational force across malicious activities such as cybercrime, surveillance, propaganda, fraud, and weapons development, enabling even small groups to conduct sophisticated campaigns at scale with less human intervention. These findings underscore a significant shift in the economics and speed of cyber threats, where AI-driven automation multiplies attackers’ capabilities and complicates defense efforts.

https://securityaffairs.com/198905/ai/anthropic-ai-misuse-is-entering-a-new-phase-from-cybercrime-to-surveillance-propaganda-and-weapons.html

The Exposed AI Supply Chain

, ,

A recent internet-wide census identified 36,769 self-hosted AI endpoints, including model servers, agent builders, and vector stores, with only 2% employing HTTP authentication, leaving the vast majority exposed without network-layer protections. Notably, 6,935 Ollama model servers responded to anonymous requests, and 5,223 agent builder endpoints—tools that store sensitive credentials—were widely reachable without authentication, posing significant security risks. The study highlights widespread misconfiguration rather than software vulnerabilities, recommending measures like binding services to localhost, deploying authentication gateways, and treating credential-storing agent builders as secure secrets stores to reduce exposure.

https://www.mysteriumvpn.com/blog/data-and-research/we-exposed-ai-supply-chain

ClickFix Attacks Infecting PCs and Macs Are Going Viral

, , , ,

ClickFix attacks have rapidly become widespread on both PCs and Macs by tricking users into pasting and executing malicious commands via fake CAPTCHA prompts on compromised websites. This social engineering exploits users' fatigue with complex web interactions, prompting them to unknowingly run malware without the need for traditional code-signing or heavy infrastructure. Security firms have observed ClickFix variants bypassing protections like macOS Gatekeeper, with attackers increasingly leveraging public services and blockchain for control, making this an enduring and challenging threat to mitigate through technology alone.

https://arstechnica.com/security/2026/09/clickfix-attacks-infecting-pcs-and-macs-are-going-viral/

Instagram Users Face Ransom Demands Over Fake Copyright Claims

, , ,

Instagram users with large followings have been targeted by scammers filing fake copyright infringement claims to suspend their accounts and then demanding ransom payments to withdraw these bogus claims. Despite Meta's assurances of protections and content restoration, creators report slow appeals processes and inadequate detection of abuse, leading some to pay extortion demands to quickly regain access, only to face repeated strikes. Experts note that automated tools and platform policies assuming claim legitimacy have enabled this scam to proliferate, worsening the challenges of distinguishing genuine copyright claims from fraudulent ones.

https://www.bbc.com/news/articles/cjw54ww73qjo

Four Spy Groups Used the Same Chrome and Windows Exploit Kit Within a Week

, ,

Multiple espionage groups, including the China-aligned APT31 and several others suspected of Chinese links, rapidly adopted a previously undocumented exploit kit named BlueMoon within a week to target Windows and Chrome vulnerabilities. BlueMoon chains three vulnerabilities in Chrome's V8 engine and Windows ALPC to achieve code execution and privilege escalation, often delivered through spear-phishing and used to deploy malware via DLL sideloading and malicious Chrome extensions. Despite patches being released, organizations are urged to check for persistent artifacts like malicious browser extensions, scheduled tasks, and suspicious files, as the exploit kit's spread may continue due to its ease of adoption and possible AI-assisted development.

https://thehackernews.com/2026/09/four-spy-groups-used-same-chrome-and.html

Chrome V8 Zero-Day Exploited in the Wild Enables Code Execution Inside Sandbox

, ,

Google released an update patching 230 security vulnerabilities in Chrome, including an actively exploited medium-severity zero-day (CVE-2026-87491) in the V8 JavaScript engine that allows remote code execution inside the sandbox via crafted HTML pages. The U.S. CISA added this vulnerability to its Known Exploited Vulnerabilities catalog, mandating federal agencies to apply patches by September 23, 2026, while users and Chromium-based browser operators are urged to update promptly to mitigate attacks.

https://thehackernews.com/2026/09/chrome-v8-zero-day-exploited-in-wild.html

Microsoft Patches Record 974 Flaws, Including Two Exploited Windows Zero-Days

, ,

Microsoft released a record-breaking Patch Tuesday update addressing 974 vulnerabilities across its software portfolio, including two actively exploited Windows zero-day flaws that allow local privilege escalation. The update covers critical issues mainly involving privilege escalation, remote code execution, and information disclosure, and prompted the U.S. Cybersecurity and Infrastructure Security Agency to mandate urgent patching for federal agencies. Despite the high volume, security experts emphasize prioritizing fixes based on exploitability and risk context to effectively reduce attack surfaces.

https://thehackernews.com/2026/09/microsoft-patches-record-974-flaws.html

The Shared Clipboard Inside the Sandbox: Cross-Account Data Leakage in ChatGPT

,

Check Point Research identified a covert cross-account communication channel in ChatGPT’s sandboxed code-execution environment that allowed attackers to execute hidden tasks using a victim’s session, including accessing connected apps like Gmail and exfiltrating data without the victim’s knowledge. This channel exploited shared access to an internal package service’s item properties, effectively turning it into a “shared clipboard” across isolated containers from different accounts. OpenAI has since decommissioned the vulnerable internal service, but the finding highlights challenges in securing AI platforms where internal shared infrastructure and broad tool access can lead to unintended data leakage.

https://research.checkpoint.com/2026/the-shared-clipboard-inside-the-sandbox-cross-account-data-leakage-in-chatgpt/

The Purpose of DNS Is to Spread Scams

, ,

A recent analysis reveals that up to 20% of newly registered generic top-level domains (gTLDs) are used for scams, highlighting a significant abuse of the Domain Name System (DNS) by criminals. Major registrars like NameCheap facilitate these registrations, but suspension rates of malicious domains remain low, allowing fraudsters to exploit the rapid activation of domains for phishing and spam campaigns. While ICANN and the industry are exploring measures like stronger customer verification and coordinated abuse reporting, striking a balance between open domain registration and fraud prevention remains a complex challenge.

https://shkspr.mobi/blog/2026/09/the-purpose-of-dns-is-to-spread-scams/

Once Popular for Attacking AI, ASCII Smuggling Is Embraced by Spammers

, , ,

A Unicode block once used to stealthily embed malicious prompts in AI attacks—known as ASCII smuggling—is now being exploited by spammers to evade email filters. This block contains invisible characters that disrupt keyword detection by machine learning–based spam classifiers, allowing spammers to obfuscate trigger words without alerting human readers. Microsoft observed a sharp surge in such obfuscated spam emails in early 2026 and recommends updated filtering techniques to counter this evasion tactic.

https://arstechnica.com/security/2026/09/once-popular-for-attacking-ai-ascii-smuggling-is-embraced-by-spammers/

Fake Software Installers Disable Windows Update and Weaken Microsoft Defender

,

A malware campaign impersonating trusted software vendors via counterfeit download sites targets primarily Chinese-speaking users and organizations, deploying malicious installers that disable Windows Update, weaken Microsoft Defender, and establish persistent access. The attacks, linked to the Chinese threat cluster Silver Fox, use scheduled tasks and PowerShell to evade detection, modify system defenses, and communicate with attacker-controlled command-and-control servers. Victims span multiple sectors including healthcare, manufacturing, and government, while the campaign employs sophisticated tactics like DLL sideloading and code-signing certificate abuse to deliver backdoors such as ValleyRAT.

https://thehackernews.com/2026/09/fake-software-installers-disable.html

You Don’t Want This Sleepwalker Backdoor on Your Windows Machine

, ,

Security researcher Dominik Reichel uncovered a sophisticated Windows backdoor called Sleepwalker that resides stealthily in memory, waiting for a specific encrypted network packet to activate its unique 23-instruction command language. Disguised as Microsoft’s dpapi.dll and loaded via side-loading into an ESET Management Agent process, Sleepwalker avoids detection by not initiating outbound traffic or listening on network ports, indicating a well-resourced targeted operation rather than opportunistic malware. While many details about its deployment and operators remain unknown, Reichel has released tools and guidance to detect and mitigate the threat.

https://www.theregister.com/security/2026/08/24/you-dont-want-this-sleepwalker-backdoor-on-your-windows-machine/5292021

Malware Can Abuse Windows Hello for Business Keys for Persistent Entra ID Access

, ,

Researcher Dirk-jan Mollema demonstrated that malware running in a logged-in Windows session can abuse Windows Hello for Business keys to authenticate silently to Microsoft Entra ID, enabling persistent cloud access without requiring admin privileges or extracting private keys. This technique leverages Windows ticketing and WebAuthn to request authentication on a compromised device, allowing attackers to register new devices, obtain refresh tokens, and bypass some conditional access controls, exposing limits in phishing-resistant authentication methods. Mollema has released proof-of-concept scripts and recommends monitoring for unexpected device registrations to detect such abuse.

https://thehackernews.com/2026/08/malware-can-abuse-windows-hello-for.html

Claude, Codex, and Hermes Installed Unowned Code Inside Corporate Networks

, ,

Researchers found that AI coding agents like Claude, Codex, and Hermes executed unowned code within corporate networks by following installation commands listed in misconfigured llms.txt files on public websites. These files, intended to guide AI agents, included references to non-existent software packages and domains that attackers could claim to deliver malware, demonstrated by proof-of-concept tests showing real Fortune 500 companies inadvertently ran such code. The findings reveal a critical security gap where AI agents treat third-party documentation as authoritative without verifying it, enabling supply-chain risks and undermining traditional boundaries between data and executable code.

https://arstechnica.com/security/2026/08/claude-codex-and-hermes-installed-unowned-code-inside-corporate-networks/

Hidden Prompts Trick AI Into False Email Summaries

, , ,

Researchers at Forcepoint X-Labs demonstrated that attackers can embed hidden HTML prompts in emails to manipulate AI-powered summarizers into producing false information without alerting users. Their proof-of-concept showed altered email summaries with incorrect financial and scheduling details, highlighting the risks of prompt injection attacks on AI systems that process untrusted external content. To mitigate these threats, organizations should separate trusted instructions from untrusted data, verify AI-generated outputs against source content, and enforce strict controls on AI assistant privileges.

https://www.darkreading.com/cyber-risk/hidden-prompts-trick-ai-false-email-summaries

Scroll to Top