Cybersecurity Roundup: Major Threats, Vulnerabilities, and Industry Shifts This week’s cybersecurity landscape saw high-profile feuds, novel attack techniques, and critical infrastructure risks, alongside policy updates and emerging malware trends. ### Clop Ransomware Gang’s Leak Site Hijacked in Feud The Cl0p ransomware gang’s Tor-based data leak site was defaced by the ShinyHunters hacking group, which claims to have stolen server logs, source code, and private keys for Clop’s onion service. The attack stems from a long-standing feud over Clop’s alleged threats during its Oracle E-Business Suite campaign. ShinyHunters demanded an eight-figure payment and a public apology, threatening to expose companies that paid ransoms to Clop. ### BragJack Flaws Exploit Browser AI Assistants Researchers at Forever disclosed BragJack, a set of vulnerabilities allowing malicious browser extensions to hijack built-in AI assistants in Chrome, Edge, Opera Neon, Perplexity Comet, and Claude in Chrome. By injecting scripts or tampering with network traffic, attackers could execute unauthorized commands such as reading emails, accessing files, or activating cameras without user interaction. Vendors awarded bounties ranging from $600 to $7,000 for the findings. ### Worm-Ready Go Implant Targets AI Agent Tooling A threat actor published malicious versions of MemTensor’s MemOS packages on npm and PyPI, embedding a Go-based implant named sckit. Unlike typical malware, it activates when the Python library is imported or the npm plugin is used, hunting for npm, PyPI, GitHub, AWS, and Hugging Face secrets. While the implant includes self-spreading templates, no active propagation has been observed. Aikido and StepSecurity also reported on the threat. ### AI Relay Networks Mask Chinese Traffic to Western Models Team Cymru identified nearly 11,000 servers running Claude Relay Service or its successor, sub2api, which pool AI accounts to bypass regional restrictions. In one U.S.-hosted cluster, 4,000+ Chinese and Hong Kong IPs accessed OpenAI, Anthropic, xAI, and Google endpoints via relays, obscuring their true origin. ### Infostealer Logs Expose Remote Access Keys in U.S. Water Sector SpyCloud analyzed stolen credentials tied to 10,000 U.S. water and wastewater utilities, finding 1,787 organizations with active infostealer exposure. Among them, 258 had credentials for OT or remote-access systems, including TeamViewer, SonicWall, and Fortinet portals. One case revealed a single compromised device at a metering technology provider exposing logins for 167 utility portals. ### CLOSEDQUORUM: AI-Powered Malware Replaces C2 Servers Cisco Talos documented CLOSEDQUORUM, a Go-based Windows implant that uses commercial LLMs (DeepSeek, Qwen, Mistral, Gemini) to decide actions such as credential theft or persistence via a voting system. The malware exfiltrates data to Discord channels but has not been confirmed in the wild. Public builds contain placeholder API keys, though custom versions suggest targeted deployment. ### Ubuntu Accelerates Kernel Fixes Amid Rising CVEs Canonical announced a shift to a single two-week kernel update cycle for Ubuntu, releasing new kernels weekly due to a surge in CVEs driven by AI-assisted bug discovery and the Linux kernel’s expanded CVE assignments. Admins can test fixes early via the -proposed pocket, with workarounds or hardening guidance promised within 24–48 hours of disclosure. ### Critical TDengine Flaw Threatens Industrial Telemetry Ridge Security disclosed CVE-2026-42542, a high-severity integer underflow in TDengine, a time-series database used in industrial telemetry, energy, and IoT. The unauthenticated flaw allows DoS via a single malformed packet, with potential for further exploitation due to heap corruption. Affected versions (3.4.0.0–3.4.1.5) are patched in 3.4.1.6. ### RemControl: Android Banking Trojan with AI Roots Group-IB uncovered RemControl, a malware-as-a-service Android trojan spreading via fake Google Play pages for the TVTap IPTV app. Targeting 30+ banks in Western Europe, the Middle East, and Canada, it abuses Accessibility permissions to overlay phishing screens, log keystrokes, and enable full remote control. Evidence suggests parts of the platform were built using an AI assistant misled into believing it was developing a quiz app. ### Docker Botnet Prioritizes AI API Key Theft ThreatDown detailed CARBONATO, a botnet that exploits unauthenticated Docker daemons (port 2375) to deploy Hermes Agent, an AI framework repurposed for malicious commands via Telegram. The malware scans networks every five minutes, prioritizing AI API key theft over other credentials. Operators are believed to be based in Costa Rica, based on infrastructure and language clues.

Critical 18-Year-Old Linux Kernel Flaw (CVE-2026-64564) Enables Root Access and Container Escapes A severe use-after-free vulnerability in the Linux kernel, dubbed SCTPhantom (CVE-2026-64564), has been disclosed, allowing attackers with unprivileged local access to escalate to root privileges or escape containers to compromise the host system. The flaw, introduced in Linux 2.6.25 (December 2007), remained undetected for nearly 18 years before its discovery. The bug resides in the SCTP Dynamic Address Reconfiguration feature, specifically in how the kernel processes ASCONF chunks a mechanism defined in RFC 5061 for managing network paths in SCTP associations. The vulnerability stems from an identity mismatch in address validation: while the kernel checks a DEL-IP operation using the packet’s source address, a cached pointer relies on the address parameter tied to the actual network path. By crafting a malicious ASCONF sequence (e.g., adding, deleting, and then wildcard-deleting an address), attackers can force the kernel to dereference freed memory, triggering a use-after-free condition. Researchers at TencentOS Security Team, leveraging their Corvus AI autonomous vulnerability research system, developed a full privilege-escalation exploit chain. The attack begins by reclaiming the freed memory via a packet socket ring buffer, leaking a kernel memory address to bypass KASLR (Kernel Address Space Layout Randomization). A second use-after-free, combined with SCTP authentication key manipulation, constructs a fake kernel object graph, ultimately executing commit_creds to grant root access all without traditional shellcode or ROP chains. The exploit also enables container-to-host escapes, bypassing default seccomp profiles by leveraging per-socket SCTP options instead of system-wide sysctls. Testing across Ubuntu 24.04, Debian 13, Rocky Linux 9, and kernels from 5.14 to 7.2, the attack succeeded in all environments, including six of eight container escape attempts. With a CVSS v4.0 base score of 8.5 (High), the flaw poses significant risk due to its low attack complexity and severe impact on confidentiality, integrity, and availability. A patch (commit 9b245f86f0b) was merged upstream, rejecting DEL-IP requests targeting transports still referenced by active ASCONF chunks. Fixes have been backported to stable branches (6.6.148, 6.12.101, 6.18.42, 7.1.6), and CVE-2026-64564 was formally announced on August 4, 2026, following private disclosure on July 12. Systems running SCTP-enabled kernels, particularly in multi-tenant or containerized environments, are urged to apply updates immediately.

Critical Linux Kernel Flaw Exposes SSH Keys and Password Hashes A newly disclosed Linux kernel vulnerability, tracked as CVE-2026-46333 and dubbed "ssh-keysign-pwn," allows attackers to extract highly sensitive data including SSH private keys and password hashes from affected systems. The flaw stems from a race condition in the kernel’s ptrace access control logic, specifically within the `__ptrace_may_access()` function. ### How the Exploit Works The vulnerability arises when a privileged process (e.g., ssh-keysign or chage) shuts down. During this brief window, its memory context is cleared (mm = NULL), but its file descriptors remain open. An unprivileged local attacker can exploit this gap using `pidfd_getfd()` to steal these descriptors, bypassing intended permission checks. A proof-of-concept (PoC) exploit on GitHub demonstrates how attackers can repeatedly spawn processes to race against a privileged helper’s exit, successfully extracting file descriptors in 100–2000 attempts making it a practical threat. ### Impact & Risks - SSH Private Key Theft: Enables attackers to impersonate systems or users, conduct man-in-the-middle (MitM) attacks, and move laterally across networks. - Password Hash Exposure: Full read access to `/etc/shadow`, allowing offline cracking of credentials. - Cascading Compromises: Since SSH keys are often reused, a single breach can lead to wider network access. ### Affected Systems The flaw impacts most Linux distributions running kernels before the May 14, 2026 patch, including: - Ubuntu - Debian - Arch Linux - CentOS - Raspberry Pi OS Given the vulnerability’s six-year existence, many long-term deployments remain exposed. ### Mitigation & Response - Apply kernel patches for CVE-2026-46333. - Rotate all SSH keys, particularly on critical systems. - Audit access to sensitive files like `/etc/shadow`. - Monitor for suspicious `ptrace` or `pidfd` system calls. - Restrict local user access where possible, as exploitation requires local presence. With a public PoC exploit already available, the risk of active exploitation in the wild is heightened, underscoring the urgency for remediation.

Dirty Frag: New Linux Kernel LPE Vulnerability Grants Root Access Across Major Distros A newly disclosed Linux kernel vulnerability, dubbed Dirty Frag, enables local privilege escalation (LPE) by chaining two page-cache write flaws xfrm-ESP Page-Cache Write and RxRPC Page-Cache Write to achieve root access on nearly all major Linux distributions. The exploit, publicly released on May 7, 2026, following an embargo break, leverages a deterministic logic flaw rather than race conditions, ensuring a high success rate without kernel panics. Discovered by security researcher Hyunwoo Kim (@v4bel), Dirty Frag exploits the kernel’s zero-copy send path, where `splice()` inserts a reference to a read-only page cache (e.g., `/etc/passwd` or `/usr/bin/su`) into the `frag` slot of a sender-side `sk_buff`. Receiver-side cryptographic operations then modify the page cache in-place, corrupting files even for unprivileged users. ### Exploit Mechanics 1. xfrm-ESP Variant: - Targets `esp_input()` in the IPsec ESP receive path, skipping buffer allocation checks (`skb_cow_data()`) for non-linear `skb`s. - Attackers use `XFRMA_REPLAY_ESN_VAL` to overwrite arbitrary bytes (e.g., `/usr/bin/su`) with a root-shell ELF, requiring user namespace creation (`unshare(CLONE_NEWUSER)`), which is blocked on some Ubuntu systems via AppArmor. 2. RxRPC Variant: - Exploits `rxkad_verify_packet_1()` to perform in-place decryption on the first 8 bytes of an RxRPC payload. - Attackers brute-force a session key to manipulate plaintext (e.g., emptying `/etc/passwd`’s password field), bypassing PAM authentication. This variant does not require namespace privileges but relies on the `rxrpc.ko` module, absent by default on RHEL but present on Ubuntu. Chaining both exploits ensures root access across distributions, with the PoC first attempting the ESP path before falling back to RxRPC if `unshare` fails. ### Affected Systems The vulnerabilities span nine years, with the ESP flaw introduced in January 2017 (commit `cac2661c53f3`) and the RxRPC flaw in June 2023 (commit `2dc334f1a63a`). Confirmed affected distributions include: - Ubuntu 24.04.4 (kernel 6.17.0-23) - RHEL 10.1 (kernel 6.12.0-124.49.1) - openSUSE Tumbleweed (kernel 7.0.2-1) - CentOS Stream 10, AlmaLinux 10, Fedora 44 ### Patches & Mitigation - The ESP patch, using `SKBFL_SHARED_FRAG` to enforce buffer isolation, was merged into the netdev tree on May 7, 2026. - The RxRPC patch remains unmerged upstream. - No CVEs have been assigned due to the premature embargo break. - Temporary mitigation involves blacklisting the affected modules (`esp4`, `esp6`, `rxrpc`) via: ```bash sh -c "printf 'install esp4 /bin/false\ninstall esp6 /bin/false\ninstall rxrpc /bin/false\n' > /etc/modprobe.d/dirtyfrag.conf; rmmod esp4 esp6 rxrpc 2>/dev/null; true" ``` This disrupts IPsec and RxRPC functionality, requiring careful evaluation for systems reliant on VPNs. The full technical write-up and PoC are available on the researcher’s GitHub repository.

High-Severity Linux Privilege Escalation Flaw "Pack2TheRoot" Disclosed Deutsche Telekom’s Red Team has publicly disclosed a high-severity privilege escalation vulnerability, CVE-2026-41651 (CVSS 8.8), dubbed Pack2TheRoot, affecting default installations of major Linux distributions. The flaw, present in the PackageKit daemon a widely used package management abstraction layer allows any local unprivileged user to silently install or remove system packages, ultimately gaining full root access without authentication. The vulnerability impacts PackageKit versions 1.0.2 through 1.3.4, spanning over 12 years of releases and exposing systems across Debian, Ubuntu, Fedora, and Red Hat-based distributions, including enterprise servers running Cockpit. Confirmed vulnerable default installations include: - Ubuntu Desktop (18.04, 24.04.4 LTS, 26.04 LTS Beta) - Ubuntu Server (22.04, 24.04 LTS) - Debian Desktop (Trixie 13.4) - Rocky Linux Desktop (10.1) - Fedora (43 Desktop and Server) Exploitation is straightforward: an attacker with basic local access can bypass authorization controls, install malicious packages, or remove critical security components. A proof-of-concept (PoC) exists, reliably achieving root code execution in seconds, though it remains undisclosed. The flaw was discovered during Telekom Security’s research into local privilege escalation vectors, with Claude Opus (Anthropic) assisting in the investigation starting in 2025. Findings were responsibly disclosed to PackageKit maintainers, who confirmed the issue and its exploitability. While the attack leaves detectable traces such as PackageKit daemon crashes logged in *journalctl* systems can be checked for vulnerability using: - Debian/Ubuntu: `dpkg -l | grep -i packagekit` - RPM-based: `rpm -qa | grep -i packagekit` - Daemon status: `systemctl status packagekit` or `pkmon` A patch was released in PackageKit 1.3.5 (April 22, 2026), with distribution-specific fixes available via: - Debian: [security-tracker.debian.org](https://security-tracker.debian.org) - Ubuntu: Launchpad CVE tracker - Fedora: PackageKit-1.3.4-3 (via Koji) Administrators are advised to apply updates immediately, particularly on internet-facing servers running Cockpit.

Critical 7-Zip Vulnerability Exposes Millions of Systems to Remote Code Execution A high-severity vulnerability (CVE-2024-, rated 8.8) in the widely used open-source utility 7-Zip has been disclosed, allowing attackers to execute malicious code simply by tricking users into opening a crafted archive. The flaw affects all versions prior to 26.01, released in late April, and requires no further interaction merely opening a booby-trapped file (.7z, .zip, .rar, or even disguised NTFS disk images) on a system with at least 16 GB of RAM* is sufficient for exploitation. The impact is severe due to 7-Zip’s ubiquity. Beyond the Windows GUI application, command-line variants across multiple operating systems including Linux distributions with outdated p7zip ports are vulnerable. Millions of CI/CD pipelines, automated scripts, and server processes that interact with archives (even just to list contents) are at risk. With over 400 million downloads on SourceForge and 24.5 million via Chocolatey, plus widespread inclusion in Linux servers, VMs, and Docker images, the potential attack surface spans hundreds of millions of machines. The vulnerability extends further due to 7-Zip’s integration into third-party software. Antivirus scanners, backup tools, log analyzers, file managers, and malware sandboxes often embed 7-Zip’s libraries, many of which run with elevated permissions. Exploitation requires no user interaction in these cases, enabling drive-by attacks via poisoned archives. The flaw stems from a bug in 7-Zip’s handling of NTFS disk images, where an attacker can manipulate buffer values to trigger arbitrary code execution. Notably, 7-Zip ignores file extensions, relying instead on file headers meaning malicious NTFS images can be hidden within standard archive formats. Testing confirms vulnerable versions are present in Ubuntu 24/26, RHEL 8, Fedora, and many OEM systems that bundle 7-Zip by default. Without built-in update mechanisms, mitigation depends on manual upgrades or package manager updates.

New Linux Kernel Vulnerability (CVE-2026-64531) Enables Local Privilege Escalation via OVSwrap Flaw A critical Linux kernel vulnerability, tracked as CVE-2026-64531 and dubbed OVSwrap, has been disclosed, allowing unprivileged local users to escalate privileges to root on a wide range of Linux distributions. The flaw resides in the Open vSwitch (OVS) kernel datapath, a networking component commonly used in cloud, container, and virtualization environments. The vulnerability was discovered by researcher Asim Viladi Oglu Manizada using an experimental approach combining large language models with structured memory-geometry visualizations to analyze complex kernel memory bugs. The issue stems from a 16-bit length field limitation in Netlink attributes, which OVS uses to store network actions. While the kernel permits action streams exceeding 64 KiB, it failed to validate whether individual nested actions such as a CLONE wrapping multiple small conntrack actions remained under the 16-bit ceiling. When an attacker crafts an action large enough to exceed 65,535 bytes, the stored length value wraps around to a small number. The kernel then misinterprets attacker-controlled data as legitimate actions, enabling exploitation without memory grooming. The attack is highly reliable, resembling a logic bug rather than a traditional memory-corruption flaw. Exploitation requires no pre-existing OVS bridge, daemon, or administrative rights an unprivileged local user can trigger the vulnerability by creating a user and network namespace (e.g., via `unshare -Urn`), gaining CAP_NET_ADMIN, and initializing a private OVS datapath. Since most distributions ship OVS as a loadable kernel module that auto-loads on demand, the attack surface exists even on systems where OVS was never explicitly installed. The flaw affects default configurations across major Linux distributions, including AlmaLinux, Debian, Fedora, Ubuntu, Rocky Linux, Arch Linux, openSUSE Tumbleweed, Amazon Linux, Kali Linux, NixOS, and Linux Mint. Some systems require minor configuration adjustments, while older kernel branches (pre-2025) remain unaffected. The vulnerable code existed for 13 years but only became exploitable after a size limit was removed in 2025. Patches have been released in stable kernel versions 5.15.212, 6.1.178, 6.6.145, 6.12.97, 6.18.40, and 7.1.5. Temporary mitigations include blacklisting the openvswitch module, disabling unprivileged user namespaces, or deploying an emergency BPF-based mitigation released alongside the proof-of-concept.