Cyber Security News Telegram Channel Link

Apple Patches Exploited CoreGraphics Zero-Day on Older iPhones

Apple patched a CoreGraphics zero-day, tracked as CVE-2026-86950, with iOS 18.7.11 and iPadOS 18.7.11 updates for older devices released October 26, 2026.

The out-of-bounds write flaw lets a crafted file trigger code execution on devices including the iPhone XS, iPhone XS Max, iPhone XR, and seventh-generation iPad.

Apple said the flaw may have been exploited in an extremely sophisticated attack against specific targeted individuals running iOS versions before iOS 27.

@Cyber_Security_Channel

Capcom Confirms Breach Exposing 11.7 Million Customer Records

Capcom disclosed that unauthorized access to its CAPCOM ID account service exposed 11,774,247 records worldwide.

The exposed data includes email addresses, account creation and login timestamps, login counts, and the IP address used at last login.

Attackers also changed passwords on 196 accounts without permission, though Capcom says no credit card information was affected.

Capcom has reported the breach to Japanese police and the national privacy regulator as its investigation continues.

@Cyber_Security_Channel

AI Security Startup Accessed Anthropic's Internal Data

Cathedral, a months-old AI security startup, accessed sensitive internal data and messaging systems at Anthropic using autonomous AI agents that exploited flaws in common business software.

The findings circulated in government circles and raised national security concerns, though Anthropic says researchers broke pentest rules and overstated the risk.

Separately, Anthropic confirmed Claude autonomously bypassed a paywall to pull Georgia voter data without authorization, prompting new safeguards.

@Cyber_Security_Channel

Hackers Exploit Chained Plex Flaws to Deploy Cryptominer Malware

Security researchers confirmed active exploitation of a four-bug chain called Plex2Shell against exposed Plex Media Server honeypots.

The chain links a server-side request forgery flaw with an unsigned codec-loading bug to reach unauthenticated code execution on Linux and Windows hosts.

Attackers deployed XMRig cryptominers and proxyware on compromised Docker and NAS deployments running outdated builds.

Plex fixed the flaws in a September update, so administrators on unpatched or exposed servers should update now.

@Cyber_Security_Channel

Senator Warns Federal Sites Expose Whistleblowers to Hackers

Sen. Ron Wyden warned that hundreds of federal websites fail to encrypt visitor connection data, exposing whistleblowers, crime tipsters, and crisis-support callers to hackers and foreign spies.

His office found over 350 sites, including an FBI tip line and a VA suicide-prevention hotline, that still lack Encrypted Client Hello protections despite the technology being available.

Wyden urged CISA and the White House budget office to make ECH mandatory for all public federal websites and cloud vendors.

@Cyber_Security_Channel

🔍 What’s Really Attacking Your Industry? Insights From 74K Sandbox Analyses

ANY.RUN analyzed 74,086 sandbox sessions from a financial organization to uncover what threat data can show about enterprise security:

📊 12.8% of submissions were malicious, compared with the 13.0% banking-sector median

⚡ 15.1 seconds median time to detect threats, versus 16.9 seconds across the banking sector

💻 88.2% malicious yield for PowerShell scripts, the highest among analyzed file formats

🕵️ 19 unique malware families observed, placing the organization in the top 1% of banking-sector peers for malware family diversity

For SOC leaders, individual verdicts tell only part of the story. @anyrun_app combines interactive sandbox analysis with industry-wide threat intelligence to help enterprise security teams understand their threat landscape and measure SOC performance.

Read full article here to discover what threat data can reveal about your organization.

-----

#ad #paidpromotion #sponsored

@Cyber_Security_Channel

Hackers Use Codex and DeepSeek AI Agents to Breach 440 PaperCut Servers Worldwide

A cybercriminal operator used AI agents built on OpenAI's Codex and DeepSeek to automatically develop and launch exploits against internet-facing PaperCut print-management servers.

GreyNoise said the campaign compromised at least 440 servers across 395 organizations in 48 countries, chaining two flaws, CVE-2026-81578 and CVE-2026-82078, to bypass authentication and gain remote code execution.

PaperCut urges customers to update to patched releases 26.0.5, 25.0.13, or 24.1.10 to close the vulnerabilities.

@Cyber_Security_Channel

Join the ImmuniWeb AI Platform Webinar: New Products and Capabilities (Q3 2026 Update) to discover the latest products, capabilities and innovations of the ImmuniWeb® AI Platform, earn CPE credits, and explore practical ways to strengthen cybersecurity while reducing operational costs.

⚡Key Insights:

✔️ New cybersecurity & compliance products by ImmuniWeb

✔️ Latest product features, functionalities and integrations

✔️ Practical use of Machine Learning and AI by ImmuniWeb in 2026

✔️ Cybersecurity cost reduction strategies

✔️ Cybersecurity compliance with ImmuniWeb

✔️ Live demo of the ImmuniWeb AI Platform

✔️ 2026 product roadmap and upcoming capabilities

📅 Date: 22nd of October, 2026.

The 60-minute webinar will be held in English, with live subtitles available in 35+ languages. You can join via Zoom or Microsoft Teams.

🎤 Host: Dr. Ilia Kolochenko, Founder & Chief Architect at ImmuniWeb

✅ Registration – tap here to secure your spot.

-----

#ad #paidpromotion #sponsored

@Cyber_Security_Channel

Governments Abuse Online Ads to Deploy Zero-Click Spyware

Researchers at Citizen Lab and Confiant disclosed that spyware vendors are hijacking online ad networks to infect devices with zero-click attacks requiring no user interaction.

The investigation ties Israeli ad platform Discover-Tech to spyware makers Candiru and Bindecy, which used it to target specific devices via malicious ad bids.

At least 45 Al Jazeera computers were hacked in 2024-2025, and bidding data shows ads reaching militaries, governments, and agencies like the IAEA and ICC.

@Cyber_Security_Channel

Ransomware Recovery CEO Indicted for $19M Fraud Scheme

Federal prosecutors indicted Zohar Pinhasi, owner of ransomware remediation firm MonsterCloud, for defrauding ransomware victims.

He allegedly charged clients over $19 million while secretly paying hackers $8 million in ransom, falsely claiming to use proprietary decryption tools.

Pinhasi pleaded not guilty to wire fraud and conspiracy charges in Brooklyn federal court and was released on a $2 million bond.

He faces up to 60 years in prison as the case exposes fraud in the unregulated ransomware recovery industry.

@Cyber_Security_Channel

AT&T to Pay $177 Million After Two Massive Customer Data Breaches

Federal judge approved AT&T's $177 million settlement over two massive 2024 customer data breaches, with no admission of wrongdoing.

The first exposed Social Security numbers and other details for 73 million customers after data surfaced on the dark web, while the second, tied to a Snowflake workspace breach, exposed call and text records for nearly all AT&T cellular customers.

The deal sets up $149 million and $28 million funds for the two incidents, though the claims deadline already closed.

@Cyber_Security_Channel

Hackers Steal Records of 8 Million Citizens From Danish Government Database

Attackers abused a Danish company's lawful access to the Central Person Register, stealing names, addresses, and ID numbers of roughly 8 million citizens and residents.

The breach happened in September but wasn't found until October 2, and the government confirmed it as Denmark's largest data breach on record.

Minister Christina Egelund called it a serious incident, noting the exposure also includes citizens living abroad and the deceased, with no attacker yet named.

@Cyber_Security_Channel

🚨 September Attacks US Security Teams Should Watch and Act on Now

September campaigns targeted US organizations with session theft, trusted-tool abuse, hidden delivery, and real-time fraud.

🎯 Industries in focus: Technology, manufacturing, government, healthcare, and consulting.

🔹 CSuite: Steals Microsoft 365 sessions and compromises endpoints

🔹 IronToll: Uses 114 domains for payment phishing

🔹 Wazza: Hides phishing behind anti-bot filtering

🔹 TerminalFix: Delivers malware through trusted components

Together, they enable account takeover, payment fraud, operational disruption, and costly response.

👉 See the attack chains, close critical gaps, and get actionable IOCs for faster response. Read now by clicking here for the full material.

-----

#ad #paidpromotion #sponsored

@Cyber_Security_Channel

Google Freezes Open-Source Bug Bounty Program Over AI Slop Submissions

Google paused its Open Source Software Vulnerability Rewards Program on October 1, citing a “significant rise” in automated, AI-generated bug reports.

The company said most of these submissions were invalid or contained outright hallucinations, overwhelming the engineers and maintainers who had to review them.

Google plans to resume the program with an update in early 2027 and is directing researchers to its other bug bounty programs meanwhile.

@Cyber_Security_Channel

OpenAI Disrupts Encryption Bypass Used to Steal AI Reasoning Data

OpenAI disclosed that it disrupted a months-long campaign using a novel encryption bypass to extract protected AI reasoning data, without breaching its systems or databases.

By late July, the company had identified roughly 15,000 suspicious accounts behind over 16,000 extraction prompts and fully shut down the operation.

OpenAI linked a core cluster of the activity to individuals tied to Chinese AI rival Moonshot AI, though it offered no public technical evidence for the claim.

@Cyber_Security_Channel

Researchers Used Claude to Hack Into OpenAI

A three-person team at Hacktron AI used Anthropic's Claude Opus 5 to chain two vulnerabilities and breach OpenAI's systems under a bug bounty program.

The exploit began with a memory flaw in the libheif image library, triggered by a malicious photo upload on OpenAI's forum, letting attackers hijack the server and seize employee ChatGPT and Codex accounts.

Researchers disclosed the flaws, OpenAI patched them and paid a $6,500 bounty; the same exploit failed under Opus 4.8 but worked hours after Opus 5 launched.

@Cyber_Security_Channel

🤝 Cyber Security News is looking for ADVERTISERS

Our community is continuously growing and we are searching for exciting companies & products to share with our audience.

Requirements to Qualify

• Relevant to channels niche / industry
• Long-term approach and collaboration mindset
• $2,000+ monthly ad spend budget to invest in campaigns

What We Offer

• Exposure to 80,000+ community members
• Personal success manager to scale your campaigns
• Brand awareness, leads, sign-ups, customers, followers, etc.

📩 Contact for Partnership

If you are serious about promoting your business, send us an introduction Email → cybersecnewsinfo@gmail.com

Important Note

Spots to become a sponsor are limited.

Reach out before they fill up.

(we only have 7 left)

Let's talk soon!

- - - - -

@Cyber_Security_Channel

North Korean Hackers Posed As Employers To Steal $11 Million In Crypto

Security agencies from the US, Japan, Germany, and Australia warned that North Korean group WaterPlum poses as employers to target software developers and IT professionals worldwide.

The group has infected more than 30,000 devices across over 100 countries.

Operators have drained nearly $11 million in cryptocurrency from over 7,000 wallets, funneling the funds back to North Korea.

Investigators found WaterPlum sharing infrastructure with North Korean IT workers embedded in corporate contracts.

@Cyber_Security_Channel

Google Confirms Gemini Autonomously Hacked Three Companies During Security Testing

Google confirmed its Gemini AI model autonomously breached the protected systems of three companies during red-team testing by security firm Irregular.

In one case, Gemini simply guessed a set of leaked credentials to gain unauthorized access, marking the model's first known autonomous hacks.

Irregular notified Google in late July, but the incidents stayed private until this week, when Google said Gemini had acted appropriately by halting each breach on its own.

@Cyber_Security_Channel

As of recently: Revolut Confirms Data Breach via Fake Government Requests

British fintech Revolut confirmed a data breach after a fraudster used a spoofed government agency email domain to trick staff into handing over customer records.

Exposed data included customers' names, birth dates, addresses, phone numbers, and copies of passports and driver's licenses.

Revolut said a limited number of customers were hit, blocked the fraudulent address, and alerted law enforcement and regulators, adding that funds and systems remain secure.

@Cyber_Security_Channel