Security Affairs AI-CYBERSECURITY newsletter includes a collection of the best articles and research on AI in the international landscape Artificial intelligence is rapidly changing cybersecurity, reshaping both the techniques used by attackers and the tools available to defenders. AI agents can automate tasks, analyze large amounts of data, discover vulnerabilities and accelerate offensive operations. At the same time, AI is creating new capabilities for threat detection, incident analysis and response. Security Affairs’ new newsletter follows this evolution, covering every week the latest threats, attacks, vulnerabilities and developments at the intersection of AI and cybersecurity. ‘It’s 100 percent a national security threat’: Accenture contractor connected an AI agent to the Pentagon’s secure network, and then sent data to a high-risk website Security notice — malicious version of Anthropic.claude-code (2.1.312 and 2.1.313) identified in the Cursor VS Code Extension Marketplace UK and Germany to jointly develop military AI in 2027 Appraisal by AI: Model-driven victim prioritization in the VOLTA infostealer AI policy wonks declare ‘sovereign agents’ a threat to civilization Progress on provable inference for large language models The Rise of LLMJacking: How Threat Actors Monetize Compromised Cloud and Model Resources “Black October” for AI Developer Platforms ARTEX Exposed: Mapping the Infrastructure Behind the South Korean Bank Breaches A thousand fewer paths to root Securing Muse: More Than $1 Million in Bug Bounties Paid Out Free the Model, Sell the Gateway: Prompt Adware hidden in Jailbreak Bundles WorkOS October 30th, 2026 Data Security Incident Covert underperformance on safety research tasks GPT‑6 Cyber: Advancing the frontier of aligned cybersecurity agents ThisBash Escape: Claude Dispatch Flaw Enables Browser Extensions RCE on the OS Follow me on Twitter: @securityaffairs and Facebook and Mastodon Pierluigi Paganini (SecurityAffairs – hacking, newsletter)
French Bank Crédit Mutuel Arkéa says hackers exploited a service provider to steal up to $133 million using the bank’s own account. French cooperative bank Crédit Mutuel Arkéa was hit by a cyberattack that let attackers make fraudulent transactions of up to $133 million, forcing it to suspend certain systems. Crédit Mutuel Arkéa is a major French cooperative banking and insurance group headquartered in Brest, Brittany, France. The bank says the attacker used its own account to send transactions to other banking organizations, and that customer assets weren’t affected. “French lender Credit Mutuel Arkea said on Thursday hackers had stolen around €100 million ($112.61 million) in a cyber heist and it was working to try to recover the money, in a rare case of cybercriminals successfully stealing from a mainstream bank.” Reuters states. Blockchain researcher ZachXBT first warned of a large unauthorized withdrawal from multiple centralized exchange accounts a day earlier. He said the attacker may have stolen up to 118 million euros ($133 million) and withdrawn it as the USDC and USDT stablecoins. “Over the past 24 hours the illicit actor has been aggressively swapping USDC for ETH, transferring to instant exchanges (WellWallet, Kyrrex, etc), converting to XMR via Wagyu, and bridging across multiple chains.” ZachXBT wrote. “While the illicit actor remains unidentified, I can say with a high degree of confidence it is NOT Lazarus Group based on multiple laundering indicators.” The attacker remains unidentified, but ZachXBT believes it is highly unlikely to be Lazarus Group, based on money-laundering patterns. According to BleepingComputer, the expert pointed out that the lack of public disclosure allowed attackers to launder $30 million by November 3, potentially preventing EU authorities from freezing the funds. The Paris prosecutor’s cybercrime unit opened an investigation after receiving a complaint from Crédit Mutuel Arkéa, with France’s cybercrime police leading the probe. Arkéa disclosed the incident on November 4, saying it was contained, customer assets were unaffected and some transfers had been delayed. Attackers reportedly accessed the bank through a service provider’s platform, not Fortuneo, and used Arkéa’s own account to move funds through other banks and convert them into cryptocurrency. The bank also warned customers about increased phishing risks. “Crédit Mutuel Arkéa announces that it has been the target of an attack that enabled the execution of fraudulent transactions.” reads the notice published by the bank. “Upon detecting this attack, and in accordance with standard procedures, the Group contacted the relevant regulatory and supervisory authorities to inform them of the situation.” “This incident is now closed and fully under control. Operations are gradually resuming this morning, following thorough verification by the Group and its subsidiaries and the implementation of enhanced security measures.” Crédit Mutuel Arkéa says customer assets were unaffected, although some transfers were delayed. The bank maintains its financial stability, warns of possible phishing attempts and urges customers not to share sensitive information. Its teams are assessing the incident and strengthening security measures, with further updates expected. Arkéa did not confirm ZachXBT’s reported figure but acknowledged significant losses and said some funds were blocked. The bank ruled out internal human error, described the incident as a sophisticated attack and is working with law enforcement to identify the attackers and establish how the breach occurred. When the FAQ asked whether 118 million euros had been stolen, Arkéa wouldn’t give a number: We do not confirm this information and can only validate the order of magnitude. Follow me on Twitter: @securityaffairs and Facebook and Mastodon Pierluigi Paganini (SecurityAffairs – hacking, Crédit Mutuel Arkéa)
Hunt.io traced ARTEX servers and AI-powered hacking tools used in breaches affecting seven South Korean financial institutions. Between late September and early October 2026, seven South Korean financial institutions had data breached through internet-facing systems. About 66,000 to 67,000 people and around 2,200 corporations were affected, and personal, employment, income and loan data was exposed. Hunt.io dug into the infrastructure behind the campaign, and its findings add a lot to what CrowdStrike and Genians already published. CrowdStrike reported on October 7 and Genians Security Center on October 12, and both pointed to ARTEX, an open-source autonomous penetration testing framework developed in China and licensed under AGPL-3.0. CrowdStrike assessed with moderate confidence that a single Chinese-speaking individual motivated by money ran it, and not an established threat group. DeepSeek v4.1 Flash was the main model, with GLM 5.3 and Grok 4.6 as backups, reached partly through a reseller domain, xcai[.]pro. The public indicators were a server at 38.244.50[.]120, which hosted the ARTEX console, plus a Hong Kong address and nine proxies. Genians said the ARTEX GitHub repository and Docker image were removed on October 11, though forks and mirrors remain. “CrowdStrike assessed with moderate confidence that the operator was a single Chinese-speaking individual motivated by financial gain rather than an established threat group.” reads the report published by Hunt.io. “The infrastructure reportedly accessed these services in part through the reseller domain xcai[.]pro. Both reports described a two-server configuration, consisting of a Hong Kong-based IP address used as primary infrastructure and 38.244.50[.]120, which hosted the ARTEX console.” Hunt.io notified the relevant national CERT before publishing, and says the report holds no personal records or credentials. The company’s own scanning showed 38.244.50[.]120 sits on Cogent’s network in Los Angeles, with 15 open ports. The ARTEX login page, with a Chinese title meaning “autonomous penetration testing console”, appears on ports 443 and 8090. Other services include a Gemini2API admin console and a Tinyproxy server. Searching its AttackCapture platform for that address, Hunt.io found another open directory at 23.248.226[.]250, a jump server in Hong Kong on AS138415. It held 42 files in 2 subdirectories, 105 MB in all, and the platform had logged warnings on September 5, September 26 and October 5. A file in it names 38.244.50[.]120 as the jump host, and its own comment tells the operator never to paste the credentials into a session. It sat in an open directory. Three scripts back up the link. They keep reverse SSH tunnels alive, run a local relay, and copy files to the jump host, one tunnel is named grok2api, and the credentials file also holds keys for Zhipu AI’s GLM models, an OpenAI-compatible relay and a proxy service. That explains how several language models ran in one campaign. The AI setup is the most striking part. A state file shows the harness is Claude Code itself, version 2.1.221, started 510 times since August 4, but it isn’t talking to Anthropic. A small proxy, ds-models-shim.py, fixes a compatibility problem by faking a model list and sending requests to DeepSeek, and another, xcai-effort-shim.py, adds a reasoning setting because the DeepSeek backend otherwise skips its reasoning step. The route runs from Claude Code through Claude Code Router, an open-source project, to xcai and then to DeepSeek. “A small local proxy, ds-models-shim.py, addresses a compatibility issue: the official /v1/models endpoint returns a 404 response when queried against a non-Anthropic backend, preventing Claude Code from discovering the available models. The shim intercepts that specific request and returns a locally defined model list containing deepseek-v4-pro and deepseek-flash. All other requests are forwarded to api.deepseek.com/anthropic.” states the report. The same file counts 19 custom skills, reusable instruction sets for the agent, with 147 recorded uses. They cover code audit, reconnaissance, web injection, WAF bypass testing, authentication testing, Telegram Mini Apps, APK download and decompiling, cloud phone control and payment logic testing. Next to the payment skill sat three self-registered test accounts with tokens for a cash-in deposit API, dated September 12, and a comment calls them an authorized, self-built test. The setup fits testing for insecure direct object reference flaws, a technique CrowdStrike and Genians described against Korean loan and inquiry services. The directory also held supporting tools. A script queried FOFA, a Chinese asset search engine like Shodan, through a third-party API, and a comment says the primary membership had expired and a backup key was in use. Proxy scripts enumerated a rotating proxy service at 154.201.79[.]246:443, where 79 unique exit IPs showed up among the first 100 identities tested. A separate directory at 23.248.226[.]250:8001 held a small staged setup for JNDI and LDAP exploitation, retrieved on October 5 at 03:42 UTC. It confirms command execution through a DNS callback, and one script lets web and LDAP traffic share a single port. Hunt.io’s analysis describes the tooling without publishing steps to reproduce it. Other recovered files include tools aimed at a cryptocurrency exchange’s admin session and withdrawals, Linux privilege escalation exploits, iOS probe source code and Android kernel and firmware utilities. Hunt.io says their connection to the Korean breaches isn’t confirmed. Hunt.io then showed how to hunt ARTEX servers. Searching for an ARTEX certificate issuer returned 19 IP addresses, and searching its malware classification returned 1,303 addresses across 30 countries, led by the United States, China and Hong Kong. Those figures show where addresses are registered, not where the operator sits. “The result returned 19 distinct IP addresses, including 38.244.50[.]120. This method is precise but limited to hosts that expose a matching certificate. Systems using a different certificate will not appear in these results.” continues the report. Matching the console’s page title returned about 1,300 IPs, and a looser search found 21 title strings, 19 of them likely ARTEX consoles, in English, Chinese, Korean, Japanese, Russian, Portuguese and Vietnamese. A malware-name search in AttackCapture gave 280 results with open directories. One at 130.94.31[.]86:18888 in Seoul held two small scripts, and another at 156.251.11[.]39 held an SMB reconnaissance toolkit whose test script writes a file named artex_probe_11876.txt. More hosts held exploitation kits for DedeCMS and FangCMS, a CAPTCHA bypass suite, a RuoYi and Spring administrative console exploit and CAPTCHA-based login automation. Together they show the framework attracting many operators. For defenders, the advice is plain. Remove public access and put administration behind VPNs or allowlists, and reset SSH passwords and rotate API keys, proxy credentials and tokens that exposed files may have revealed. Patch internet-facing services, watch for unexpected LDAP and JNDI activity, DNS callbacks and remote script downloads, and store secrets in a managed vault with least-privilege permissions. Also restrict access to model gateways, watch their requests, and enforce compliance rules on managed phones. Hunt.io’s lesson is that an investigation shouldn’t stop at the first list of indicators, because The company also published indicators of compromise in its analysis. “The findings highlight risks associated with exposed administrative services, reusable credentials, remote access tunnels, proxy infrastructure, and exploitation tooling.” concludes the report. Follow me on Twitter: @securityaffairs and Facebook and Mastodon Pierluigi Paganini (SecurityAffairs – hacking, ARTEX)
French cooperative bank Crédit Mutuel Arkéa was targeted in a cyberattack that enabled fraudulent transactions of up to $133 million, forcing it to respond by suspending certain systems. The bank says that the attacker used its own account to initiate transactions to other banking organizations, but customer assets remain unaffected. Crédit Mutuel Arkéa is a major French banking group with more than 5.8 million members and 211.5 billion euros in assets. Reports about the incident emerged on November 2, initially from blockchain fraud investigator ZachXBT, who warned of a large-scale unauthorized withdrawal from multiple centralized exchange accounts occurring a day earlier. According to the investigator, the attacker may have stolen up to 118 million euros ($133 million) and withdrawn it as USDC and USDT stablecoins.
Scammers are impersonating major brands and celebrities to lure consumers into fake gift card giveaways that collect personal information and trigger unwanted subscriptions. CTM360 reveals how these campaigns use deceptive verification steps and affiliate networks to generate revenue without delivering the promised rewards.
Passwords and API keys that developers typed into commands they approved in Claude Code are sitting in public GitHub repositories, where anyone can read them. When a user selects "Yes, and don't ask again," Anthropic's AI coding assistant can save the command, word-for-word, including the secret, in a project settings file. Software security company Apiiro said yesterday that 1,122 public
WorkOS says an AI-powered hacking agent that breached its internal sales application and accessed customer information was operated by a security researcher who claims the tool escaped its testing environment and unintentionally attacked the company.
Layer3 found residential proxy traffic using the White House network, raising concerns about potential exposure of internal systems. Layer3 Intel says it found residential proxy traffic leaving through the White House network. The researchers spotted it while browsing their new Residential Proxy Observatory, a tool that gives a global view of residential proxy and anonymized internet activity. “While our team was looking through it, we noticed residential proxy exit activity on AS2650. AS2650 is registered to the Executive Office of the President, the federal agency that hosts the White House network.” reads the report published by Layer3. A residential proxy service allows customers to use everyday devices to route their internet traffic. This makes their activity look as if it comes from the device’s normal home internet connection. Layer3 detected this activity on AS2650, a network registered to the Executive Office of the President, the agency responsible for the White House network. Layer3 has not accused anyone at the White House of wrongdoing and says its data cannot identify who or what caused the activity. Between October 5 and November 3, Layer3 detected 236 connections through this network. The traffic came from six IP addresses and was recorded on 26 of those 30 days. A total of 28 residential proxy services were able to use the network. To check their findings, researchers compared the IP addresses with data from other security services, including Spur and Synthient. Both services also identify these addresses as residential proxy exit points. Getting this data took no special skill. Anyone can sign up to a public residential proxy service, pay a few dollars and send requests through it. Layer3 says it didn’t target the network or try to steer its probes there, and it caught the activity by chance, simply by requesting global rotating access through these services. The timing suggests something more than a guest network. About half the activity falls between 5 p.m. and midnight Eastern time, and almost none happens in the early morning. Layer3 reads that as a person, not public guest Wi-Fi: someone regularly on the White House network late into the evening whose device carries a proxy SDK, a software kit that turns a device into an exit point. “Past research has shown that poorly secured proxy SDKs can allow internal network access and lateral movement. This is how past botnets have spread. In A Broken System Fueling Botnets, the researchers at Synthient documented how the Kimwolf botnet reached at least 2 million devices, mostly Android devices with an exposed ADB service, through residential proxy networks.” continues the report. “If an SDK on this network has the same kind of flaw, anyone buying proxy access could use it to reach internal systems.” That matters because poorly secured proxy SDKs have let attackers into internal networks before. Synthient’s research, A Broken System Fueling Botnets, documented how the Kimwolf botnet reached at least 2 million devices, mostly Android devices with an exposed ADB service, through residential proxy networks. If an SDK on the White House network has the same kind of flaw, anyone buying proxy access could reach internal systems. Layer3 doesn’t claim that’s the case here, and it describes the risk as conditional. The White House isn’t the only surprise. The observatory also finds residential proxies in some of the most isolated places on the planet. Saint Helena, Ascension and Tristan da Cunha, a South Atlantic territory that includes the most remote inhabited island in the world, shows 413 addresses and more than 32,000 observations, split between Starlink and the local provider, Sure Saint Helena. Layer3 also detected three instances of proxy activity in Antarctica. All came from a Cloudflare WARP address linked to McMurdo Station. The researchers jokingly noted that penguins are not suspects. On Pitcairn Island, they found one Starlink address and one observation. Norfolk Island had one Norfolk Telecom address with four observations, while Svalbard had one Starlink address in Longyearbyen with three observations. The researchers also found 388 addresses in the Falkland Islands, 43 in the Chagos Islands, 21 in Tokelau and seven in South Georgia. These findings suggest that residential proxy networks now operate in almost every corner of the world. For security teams, the main concern is that devices connected to protected networks could unknowingly become proxy exit points, potentially exposing those networks to outside traffic. However, Layer3 says its data only shows where proxy traffic exits. It cannot identify the device being used or the person behind it. As the company explains, “Our data shows the exits, not the device or the person behind them.” “Residential proxies have infected every corner of the world, and it looks like there is no place they cannot reach. Layer3 Intel tracks residential proxy and anonymization infrastructure worldwide. Even when the proxies hide in networks on the farthest corners of the planet, we can still observe them, leaving threat actors no place to hide.” concludes the report. Follow me on Twitter: @securityaffairs and Facebook and Mastodon Pierluigi Paganini (SecurityAffairs – hacking, proxy traffic)
Unit 42 examines recent cyberattacks targeting European refrigeration systems and outlines defense and mitigation strategies. The post Turning Up the Heat: Inside the Attacks on Europe’s Refrigeration System appeared first on Unit 42.
OpenAI has revealed GPT-6 Cyber, a new artificial intelligence model that has discovered 81 previously unknown high-severity Firefox vulnerabilities and developed working exploits for 948 distinct Linux kernel flaws. The company announced the model on Thursday, saying it significantly improves on earlier AI models in finding software vulnerabilities, validating their impact, and helping developers create fixes. During security testing, GPT-6 Cyber discovered 81 previously unknown high-severity vulnerabilities in Firefox and helped develop patches, many of which Mozilla has already released. In separate Linux kernel research, the model identified 2,308 potential security flaws, successfully reproduced 1,724 vulnerabilities, and developed local privilege escalation exploits for 948 distinct vulnerabilities that could allow an unprivileged user to gain root access.
Confirm this action.
Leaving now will discard your changes.