Amir Barati, 40, an Iranian who also holds Turkish citizenship, has been extradited from Montenegro to the United States to face criminal charges, police in Montenegro and a person familiar with the case told CNN on 1 October 2026 (US time). CNN reports he is accused of being part of a hacking scheme that breached more than 150 US universities and companies while targeting data worth over $3 billion, and is expected to face report and computer fraud charges in the US Southern District of New York, where he and 16 other Iranian nationals were charged in a superseding indictment in August; Montenegrin police arrested him in June at the FBI's request. According to The Record, Barati was arrested on 25 June in the coastal municipality of Kotor, a Montenegrin court issued a final extradition decision that week, and the 14-count indictment alleges the campaign was run through the Mabna Institute on behalf of Iran's Islamic Revolutionary Guard Corps; the Justice Department accuses the group of breaching 144 US universities, 42 US companies, 178 foreign universities and at least 11 foreign companies and stealing at least 31 terabytes of information. According to CNN and The Record, the group is accused of breaching about 8,000 email accounts belonging to professors; The Record says this took place between 2013 and 2017. Accused Iranian hackers rarely see a US courtroom, CNN notes. (sources: CNN; The Record) Correction (25 Oct 2026): the card now rests on CNN and The Record and uses CNN's figure of data worth over $3 billion; details that only an unreadable report carried (agency and NGO victim counts and reward offers) and a dead AP/ABC News link have been removed. (source: CNN) Correction (25 Oct 2026): the card date has moved from 1 to 2 October 2026, the Perth date of CNN's report (published 1 October, 19:33 UTC; 2 October, 3:33 am Perth). (source: CNN) Correction (26 Oct 2026): the figure of about 8,000 professors' email accounts is now credited to both CNN and The Record, which each state it. (source: CNN; The Record)
CNN — Iranian accused of involvement in $3 billion hacking scheme extradited to US (1 Oct 2026, US Eastern)
GitLab AI Gateway critical patch release (2 October 2026; Docs advisory) remediates CVE-2026-90970 (CVSS 3.1 9.9; AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H): improper neutralization in a custom flow prompt template that, under certain conditions, lets an authenticated user with Duo Agent Platform access escape the prompt-template sandbox via a crafted flow configuration and execute arbitrary commands on the AI Gateway. Impacted: AI Gateway from 18.1.6 before 19.2.4, 19.3 before 19.3.2, and 19.4 before 19.4.1. Fixed: 19.2.4, 19.3.2, 19.4.1. GitLab-hosted AI Gateways (GitLab.com, Dedicated, Self-Managed using GitLab-hosted gateway) already patched — no action. Self-Hosted AI Gateway on Self-Managed: upgrade immediately; GitLab says it conducted targeted pre-disclosure outreach to those customers. Credit: invisiblemeerkat. (sources: GitLab Docs AI Gateway patch release; BleepingComputer)
GitLab Docs — AI Gateway critical patch release 19.2.4 / 19.3.2 / 19.4.1 (CVE-2026-90970; 2 Oct 2026)
Google Android security blog states that in Android 17, enabling Advanced Protection automatically restricts AccessibilityService access exclusively to verified applications categorised as Accessibility Tools — closing a primary conduit used by banking trojans and spyware for screen reading, overlay fraud, silent permission grants and blocking uninstall, while preserving legitimate assistive tech. Advanced Protection also disables WebGPU in that mode to shrink browser attack surface. Google frames AccessibilityService abuse as a leading fraud/scam vector because the API can interact with the UI without root. Not a CVE; platform hardening for users who opt into Advanced Protection on Android 17. (source: Google security blog)
The Wall Street Journal (2 October 2026) and The Hacker News report that OpenAI parted ways with three safety-team researchers — Jasmine Wang, Tomek Korbak, and Mikita Balesni — after an internal investigation found they violated policies on accessing and handling sensitive company information. An OpenAI spokesperson said: "We have parted ways with three individuals for violating our policies on accessing and handling sensitive company information… Our investigation confirmed that these individuals mishandled sensitive information outside established channels." Reporting (WSJ/Bloomberg via THN) says the researchers shared confidential information with a third-party AI-safety organisation; Bloomberg characterised the material as relating to OpenAI infrastructure architecture (organisation unnamed publicly). Context reports note prior researcher concerns about AI development pace, NYT reporting on safety-evaluation trade-offs, and OpenAI's concurrent agent-safety posture (GPT-6.1 Astra shelved; tool-use pause after DNS filter gap; Transluce US/Canada gov-agent probes already on Cyberstack). Not a product CVE. (source: WSJ)
Wall Street Journal — OpenAI fires researchers for allegedly sharing information with AI safety group (2 Oct 2026)
The Verge (1 October 2026; Tom Warren) and BleepingComputer (2 Oct) report that unknown attackers hijacked the official Microsoft account on X (@Microsoft; ~13M followers) in a crypto pump-and-dump promoting a $Clippy token. The compromise began when @Microsoft followed and reposted a now-suspended impersonation account (@clippymsftcto) posing as Clippy; a related @ClippyMSFT account continued promoting the token with claims of a liquidity pool paired with $MSFT. Microsoft spokesperson Brent Colburn confirmed unauthorised access and posts that did not come from Microsoft; the account was secured, unauthorised posts removed, and investigation continues. Microsoft also stated it does not authorise or endorse any Clippy/$MSFT-linked cryptocurrency and will pursue legal action to remove the token materials. Distinct from the June 2024 @MicrosoftIndia crypto-drainer hijack. No product CVE. (sources: The Verge; BleepingComputer) Correction (27 Oct 2026): the card date has moved from 1 to 2 October 2026, the Perth date of The Verge's report (1 October US time; 7:04 am 2 October Perth). (source: The Verge)
The Verge — Microsoft confirms X account compromised / Clippy crypto posts (1 Oct 2026)
Wordfence CNA CVE-2026-19660 (published 2 October 2026; disclosed 1 Oct; CVSS 3.1 9.8 Critical AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H): unauthenticated authentication bypass in DiviEngine’s Divi Membership WordPress plugin through 2.3.0. The process_paypal_callback function (hooked to init) accepts a base64-encoded paypal_param GET value with no PayPal IPN validation, signature check, ownership check, or nonce, then passes an attacker-controlled user ID to wp_set_current_user() / wp_set_auth_cookie() — full site takeover including administrators. The PayPal gateway class is instantiated on every front-end request even when PayPal is not enabled, so the hook is always registered. Finder credit: 0xd4rk5id3. Divi Engine changelog still lists 2.3.0 (1 Sep 2026) as newest at check — no patched release documented. (sources: CVE.org; Wordfence CNA)
Docker published two Moby security advisories on 1 October 2026, with the CVE records following on 7 October; both are fixed in Docker Engine 29.8.2. CVE-2026-92543 (GHSA-7cfq-22r6-qp73, CVSS 4.0 score 7.6 High): Docker Engine always treats 127.0.0.0/8 and ::1/128 as insecure registry ranges, and it decides a registry is insecure if any one of the addresses its hostname resolves to falls in those ranges. Because the connection then re-dials the hostname rather than the matching address, an attacker who can shape DNS answers (one loopback address plus one of their own) can make the daemon skip certificate verification and allow plain-HTTP fallback, exposing registry credentials sent in X-Registry-Auth or substituting the image behind a trusted tag. Removing entries from insecure-registries does not help, and a trusted CA does not mitigate it. CVE-2026-92542 (GHSA-6m9p-4h64-m6vh, CVSS 4.0 score 6.9 Medium): on Linux Swarm nodes the firewall rules that mark VXLAN traffic for IPsec encryption also match datagrams forged by ordinary user processes, so an unprivileged local user can inject Ethernet frames into an encrypted overlay network on another node. No exploitation has been reported. (sources: Moby GitHub security advisories; listed on Tenable's newest CVE feed)
Moby (Docker) — GHSA-7cfq-22r6-qp73: insecure-registry fallback via malicious DNS responses, CVE-2026-92543 (1 Oct 2026)
Dell's DSA-2026-324 (1 Oct 2026) fixes five flaws in Dell System Update, the command-line tool admins use to push BIOS, firmware and driver updates to PowerEdge servers on Linux and Windows. The critical one, CVE-2026-86360 (CVSS 3.1 9.6), is a path traversal that an unauthenticated remote attacker could use for filesystem access; Dell says it can be leveraged to run arbitrary code with root privileges and fully compromise the application and the underlying operating system. The same release fixes CVE-2026-86361 and CVE-2026-86362 (both 8.2, local low-privileged elevation of privilege), CVE-2026-63697 (7.6, improper certificate validation leading to remote execution by a high-privileged attacker) and CVE-2026-71168 (7.3, local path traversal leading to code execution). Dell has not flagged any of them as exploited. Reporters credited by Dell: Ori Gabriel, saltedfish, and Nir Yehoshua of Cipher Security Labs. (source: BleepingComputer)
Dell — DSA-2026-324: Security update for Dell System Update (DSU) vulnerabilities (1 Oct 2026)
Nozomi Networks Labs (1 October 2026) describes Cling, a botnet it found while investigating a spike in exploitation of CVE-2021-35394, the Realtek Jungle SDK 'UDPServer' diagnostic-daemon command injection (CVSS 3.1 9.8; on the CISA KEV catalog since December 2021). The attack sends UDP datagrams starting with orf; plus shell commands that use BusyBox wget to fetch and run the bot. The analysed MIPS sample also carries exploits for CVE-2014-8361 (Realtek SDK), CVE-2023-26801 (LB-LINK), CVE-2024-3721 (TBK DVR), CVE-2025-34037 (Linksys), CVE-2016-10372 (Eir D1000), CVE-2023-41011 (FiberHome / China Mobile router) and CVE-2016-20016 (MVPower DVR). It persists by copying itself to /root/.cling and /usr/local/bin/.cling, adding entries to /etc/inittab, /etc/init.d/rcS and /etc/rc.d/rc.boot, and replacing the wget binary so it re-runs whenever wget is called. Its command channel imitates STUN, the NAT-traversal protocol used by Teams, Zoom, Webex and WebRTC: every ~5 seconds it queries 13 hard-coded public STUN servers (including a Google address), sends a non-standard registration packet with its mapped ports and infection tag, then takes commands hidden in the 12-byte STUN transaction ID. Nozomi saw commands appear to come from Google STUN infrastructure and assesses spoofed source addresses as the likely explanation. Capabilities: spreading, proxying, tunnelling and DDoS; observed attack orders targeted an internet provider, university infrastructure and gaming services. No infection count or confirmed outages published; no Australian victims reported. UPDATE (FortiGuard Labs, 5 Oct 2026): Fortinet tracks the same malware as ClingSTUN, a back-connect proxy backdoor, across three phases with different download servers. It was first delivered through CVE-2022-36553 (Hytec Inter HWL-2511-SS routers), then CVE-2025-34035 (EnGenius IoT cloud service) and CVE-2024-23625 (D-Link UPnP), then CVE-2021-35394 (Realtek), CVE-2023-1389 (TP-Link Archer AX21), CVE-2024-7029 (AVTECH AVM1203), CVE-2024-10915 (D-Link) and a goform buffer overflow across several vendors; SecurityWeek counts about two dozen exploited flaws including Ivanti, Lantronix, Linear, MeiG, Sunhillo and Tenda devices. Builds exist for ARM, x86, x86-64, MIPS and PowerPC. Newer versions kill the watchdog and competing malware, hide their process details behind copied PID 1 metadata, and send STUN binding requests to 24 (later 13) public endpoints. Fortinet rates it High severity.
Nozomi Networks Labs — A STUNning disguise: Cling malware masquerades as Google (1 Oct 2026)
Fortra Product Security published a 1 October 2026 BoKS (Core Privileged Access Manager) advisory batch covering eight flaws, three Critical. CVE-2026-79901 (FI-2026-012; CVSS 3.1 9.9; AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H): in deployments using BoKS keytab management, boks_keytabmd generates Active Directory service-account passwords from a predictable PRNG seeded with the current Unix timestamp — an attacker who knows the SPN and can estimate password-change time can reproduce a limited candidate set and verify offline (standard AD account can request a service ticket; admin access to BoKS/host/keytab not normally required). Fix: upgrade active BoKS Master to boks-server 9.0.0.7, restart BoKS, then rotate all affected/uncertain service-account passwords via BoKS keytab management and rebuild keytabs after max service-ticket lifetime + clock skew — installing the update alone does not secure passwords generated by an affected release. CVE-2026-12627 (FI-2026-017; CVSS 9.8; AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H): stack buffer overflow in boks_autoregisterd — remote attacker with network access to the autoregistration service (default port 6507) may trigger memory corruption; affected boks-server 8.1.0.0–8.1.0.23 and 9.0.0.0–9.0.0.6; fixed 8.1.0.24 / 9.0.0.7; workaround restrict/disable autoregistration until patched. CVE-2026-79898 (FI-2026-015; CVSS 9.1; AV:N/AC:L/PR:H/UI:N/S:C/C:H/I:H/A:H): authenticated command injection in crlserver via BCC / WSI REST or SOAP / cacrl — shell commands run as root on the BoKS Master; same fixed builds. Also patched: CVE-2026-79896 (High 7.5) unauth boks_portmux TLS ClientHello OOB read DoS, plus further High/Medium heap/OOB/temp-file/password issues. Fortra has not reported exploitation in the wild. (sources: Fortra FI-2026-017; Fortra FI-2026-012; SecurityWeek)
ASUS product security notices (last updated 1 October 2026) address two authenticated remote flaws in consumer/prosumer router firmware reachable via the web management interface. CVE-2026-14157 (Critical, CVSS v4.0 9.4): a specially crafted VPN client configuration file uploaded through the management UI can cause arbitrary command execution on firmware series 3.0.0.6_102. Until patched, ASUS says import VPN client configs only from trusted, verifiable sources — avoid publicly shared or unknown-origin OVPN/profile files. CVE-2026-13313 (High, CVSS v4.0 8.9): authenticated attacker can bypass checks (debug code left active) to enable Telnet and run commands with elevated/root privileges; affects 3.0.0.4_386, 3.0.0.4_388, and 3.0.0.6_102 series. ASUS also recommends a strong unique admin password (≥10 chars, upper/number/symbol mix) and against running untrusted scripts/tools/commands on LAN hosts that can reach the router UI. Neither router bulletin states exploitation in the wild. Separate same-day board notice: CVE-2026-93495 (CVSS v4.0 7.0) improper BIOS init — physical crafted device → memory R/W on listed Z390/C246 boards (BIOS 2203 / 1502). (sources: CyberInsider; ASUS Product Security Advisory) ASUS security-advisory HTML table was empty on fetch (likely JS) — switch primary_url when the specific bulletin permalink is listed.
CyberInsider — ASUS warns of critical router flaw via malicious VPN files (1 Oct 2026)
GitHub Security Advisory GHSA-rvm3-566m-v7fv / CVE-2026-103922 (published 1 October 2026; GitHub CNA CVSS 3.1 9.3 Critical; CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:C/C:H/I:H/A:N): Capacitor Android/iOS WebView navigation guard validated host and scheme but not path. A victim who opens an untrusted link inside an affected app can navigate a frame to /_capacitor_http_interceptor_; the native HTTP proxy fetches attacker-chosen remote content and returns it as a document at the application origin, giving same-origin access to localStorage, cookies, and registered Capacitor plugin APIs (device data, tokens, files — plugin-dependent). Disabling CapacitorHttp does not mitigate on vulnerable builds because the proxy path was still served. Affected: >=6.0.0 <6.2.2, >=7.0.0 <7.6.9, >=8.0.0 <8.3.5, >=8.3.5 <8.4.3, >=8.5.0 <8.5.1. Fixed in 6.2.2, 7.6.9, 8.3.5, 8.4.3, 8.5.1 — rebuild and redistribute mobile apps. (sources: Cyber Security News; Ionic/GitHub GHSA-rvm3-566m-v7fv)
