n8n is a workflow automation platform. From 1.77.0 to before 1.98.2, a stored Cross-Site Scripting (XSS) vulnerability was identified in n8n, specifically in the Form Trigger node's HTML form element. An authenticated attacker can inject malicious HTML via an <iframe> with a srcdoc payload that includes arbitrary JavaScript execution. The attacker can also inject malicious Javascript by using <video> coupled <source> using an onerror event. While using iframe or a combination of video and source tag, this vulnerability allows for Account Takeover (ATO) by exfiltrating n8n-browserId and session cookies from authenticated users who visit a maliciously crafted form. Using these tokens and cookies, an attacker can impersonate the victim and change account details such as email addresses, enabling full control over the account—especially if 2FA is not enabled. Users should upgrade to version >= 1.98.2.
A VBS-based infostealer called Cmimai Stealer has emerged, targeting Windows systems since June 2025. It collects system information, browser metadata, and screenshots, exfiltrating data via Discord webhooks. The malware uses PowerShell scripts for browser data collection and screen capture, running in a persistent loop every hour. It leverages WMI for system information gathering and employs JSON formatting for data exfiltration. While lacking advanced features like encrypted communication or credential theft, Cmimai Stealer serves as both an infostealer and a reconnaissance tool. Defensive considerations include monitoring high-risk process combinations, watching for specific PowerShell scripts and image files, and detecting Discord traffic with a unique User-Agent.
A new ransomware family called Charon has been identified, targeting the Middle East's public sector and aviation industry. The attack employs sophisticated APT-style techniques, including DLL sideloading, process injection, and anti-EDR capabilities. Charon uses a multistage payload extraction technique and a hybrid cryptographic scheme combining Curve25519 with ChaCha20 cipher. The ransomware exhibits network propagation capabilities and includes a dormant anti-EDR component. The campaign demonstrates a concerning trend of ransomware operators adopting APT-level techniques, posing an elevated risk to organizations. Defending against Charon requires a multilayered approach, including hardening against DLL sideloading, limiting lateral movement, strengthening backup capabilities, and reinforcing user awareness.
SocGholish, operated by TA569, functions as a Malware-as-a-Service vendor, employing deceptive 'fake browser update' lures to compromise systems. It leverages Traffic Distribution Systems like Parrot TDS and Keitaro TDS to filter and redirect victims. TA569 acts as an Initial Access Broker, enabling other cybercriminal groups to conduct follow-on attacks, including ransomware deployments. SocGholish utilizes domain shadowing and frequent domain rotation to evade detection. The malware's infection chain involves multiple stages, from compromised websites to on-device payload delivery. Notable customers include Evil Corp and MintsLoader operators. SocGholish's sophisticated filtering mechanisms and tracking techniques ensure only high-value targets receive the final payload.
Medtronic MyCareLink Patient Monitor has an internal service that deserializes data, which allows a local attacker to interact with the service by crafting a binary payload to crash the service or elevate privileges.
This issue affects MyCareLink Patient Monitor models 24950 and 24952: before June 25, 2025
The Octalyn Forensic Toolkit, a publicly available GitHub project, presents itself as a research tool but functions as a sophisticated credential stealer. It consists of a C++ payload module and a Delphi-based builder interface, allowing even low-skilled actors to generate functional binaries. The toolkit extracts browser data, Discord and Telegram tokens, VPN configurations, gaming account data, and cryptocurrency wallet artifacts. It establishes persistence, organizes stolen data, and exfiltrates it via Telegram. The malware's modular design, ease of use, and active exfiltration capability pose significant risks if misused. It employs obfuscation techniques, Windows persistence methods, and structured data theft, demonstrating a deliberate effort to evade detection and maximize impact.
A malicious CHM file targeting Poland was discovered, containing an infection chain that drops and executes a C++ downloader. The CHM file displays a decoy image while executing obfuscated JavaScript to extract and run a DLL. The downloader retrieves a payload from a domain associated with previous Belarus-linked threat activities. The payload is encrypted and appended to an image file. The infection process involves multiple stages, including the use of LOLbins and the creation of a scheduled task for persistence. This campaign shows similarities to activities attributed to threat actors like FrostyNeighbor and UNC1151, known for targeting Eastern European countries.
A Protection Mechanism Failure vulnerability in kernel filter processing of Juniper Networks Junos OS allows an attacker sending IPv6 traffic destined to the device to effectively bypass any firewall filtering configured on the interface.
Due to an issue with Junos OS kernel filter processing, the 'payload-protocol' match is not being supported, causing any term containing it to accept all packets without taking any other action. In essence, these firewall filter terms were being processed as an 'accept' for all traffic on the interface destined for the control plane, even when used in combination with other match criteria.
This issue only affects firewall filters protecting the device's control plane. Transit firewall filtering is unaffected by this vulnerability.
This issue affects Junos OS:
* all versions before 21.2R3-S9,
* from 21.4 before 21.4R3-S11,
* from 22.2 before 22.2R3-S7,
* from 22.4 before 22.4R3-S7,
* from 23.2 before 23.2R2-S4,
* from 23.4 before 23.4R2-S5,
* from 24.2 before 24.2R2-S1,
* from 24.4 before 24.4R1-S2, 24.4R2.
This is a more complete fix for previously published CVE-2024-21607 (JSA75748).
The article examines a malware variant associated with the SLOW#TEMPEST campaign, focusing on advanced obfuscation techniques used by the threat actors. The malware is distributed as an ISO file containing multiple files, including two malicious ones. The loader DLL, zlibwapi.dll, decrypts and executes the embedded payload, which is appended to another DLL. The analysis reveals sophisticated anti-analysis techniques, including Control Flow Graph (CFG) obfuscation using dynamic jumps and obfuscated function calls. The researchers demonstrate methods to counter these techniques using emulation and code patching. The loader DLL also employs an anti-sandbox check, only executing its payload if the target machine has at least 6 GB of RAM. The study highlights the importance of combining advanced dynamic analysis with static analysis to effectively understand and mitigate modern malware threats.
Check Point Research uncovered a malware campaign exploiting expired Discord invite links to redirect users to malicious servers. The attackers use a combination of techniques including ClickFix phishing, multi-stage loaders, and time-based evasions to deliver AsyncRAT and a customized Skuld Stealer targeting crypto wallets. The campaign leverages trusted cloud services for payload delivery and data exfiltration to avoid detection. The operation continues to evolve, with threat actors now able to bypass Chrome's App Bound Encryption using adapted tools like ChromeKatz to steal cookies from new Chromium browser versions. The campaign highlights how subtle features in Discord's invite system can be exploited as attack vectors.
FreshRSS is a self-hosted RSS feed aggregator. Prior to version 1.26.2, it's possible to run arbitrary JavaScript on the feeds page.
This occurs by combining a cross-site scripting (XSS) issue that occurs in `f.php` when SVG favicons are downloaded from an attacker-controlled feed containing `<script>` tags inside of them that aren't sanitized, with the lack of CSP in `f.php` by embedding the malicious favicon in an iframe (that has `sandbox="allow-scripts allow-same-origin"` set as its attribute). An attacker needs to control one of the feeds that the victim is subscribed to, and also must have an account on the FreshRSS instance. Other than that, the iframe payload can be embedded as one of two options. The first payload requires user interaction (the user clicking on the malicious feed entry) with default user configuration, and the second payload fires instantly right after the user adds the feed or logs into the account while the feed entry is still visible. This is because of lazy image loading functionality, which the second payload bypasses. An attacker can gain access to the victim's account by exploiting this vulnerability. If the victim is an admin it would be possible to delete all users (cause damage) or execute arbitrary code on the server by modifying the update URL using fetch() via the XSS. Version 1.26.2 has a patch for the issue.
A sophisticated campaign using typo-squatted 'Spectrum' domains has been uncovered, spreading a new Atomic macOS Stealer (AMOS) variant. The attack, disguised as a CAPTCHA verification, employs dynamic payloads based on the victim's operating system. For macOS users, a malicious shell script steals system passwords and downloads an AMOS variant. The script uses native macOS commands to harvest credentials, bypass security, and execute malicious binaries. Russian-language comments in the source code suggest involvement of Russian-speaking cybercriminals. The campaign's delivery sites show flawed logic, indicating hasty assembly. This multi-platform social engineering attack targets both consumer and corporate users, highlighting an increasing trend in cross-platform threats.
A threat actor campaign observed in late 2023 and early 2024 targeted multiple regions, initially in the Far East and later Sweden, using DLL sideloading techniques combined with the Minhook library to intercept Windows API calls. The attackers leveraged compromised digital signatures, including an expired signature from a Korean game developer, to sign components and evade detection. The final payload deployed was Cobalt Strike, a known post-exploitation tool. Three distinct sideloading scenarios were identified: MiracastView, PrintDialog, and SystemSettings. The clean loader used in the attack was sourced from infected systems rather than bundled with the sideloading package, complicating detection. Indicators include multiple file hashes and suspicious domains. This campaign demonstrates sophisticated use of API hooking and sideloading to maintain persistence and evade security controls, posing a medium severity risk to affected organizations, especially in Sweden.
A vulnerability in binary-husky/gpt_academic version 310122f allows for a Regular Expression Denial of Service (ReDoS) attack. The application uses a regular expression to parse user input, which can take polynomial time to match certain crafted inputs. This allows an attacker to send a small malicious payload to the server, causing it to become unresponsive and unable to handle any requests from other users.
A maliciously crafted QPY file can potential execute arbitrary-code embedded in the payload without privilege escalation when deserialising QPY formats < 13. A python process calling Qiskit 0.18.0 through 1.4.1's `qiskit.qpy.load()` function could potentially execute any arbitrary Python code embedded in the correct place in the binary file as part of specially constructed payload.
An issue in Casa Systems NTC-221 version 2.0.99.0 and before allows a remote attacker to execute arbitrary code via a crafted payload to the /www/cgi-bin/nas.cgi component.
CVE-2024-35944 is a medium severity vulnerability in the Linux kernel's VMware VMCI driver where a memcpy() operation in the dg_dispatch_as_host() function causes a run-time warning due to copying data across multiple structure members. This issue was detected by the Syzkaller fuzzer and relates to copying a vmci_datagram structure plus additional payload bytes into a combined structure, triggering a FORTIFY_SOURCE warning. The vulnerability has been fixed in multiple Linux kernel versions starting from 5.4.0-1075.80~18.04.1 and others. A patch is available that corrects the memcpy() usage by splitting the copy into two parts to avoid the warning. No known exploits in the wild have been reported. Users should update to the fixed kernel versions and reboot to apply the fix.
Due to lack of proper memory management, when a victim opens a manipulated Right Hemisphere Binary (.rh, rh.x3d) file received from untrusted sources in SAP 3D Visual Enterprise Viewer - version 9, it is possible that a Remote Code Execution can be triggered when payload forces a stack-based overflow or a re-use of dangling pointer which refers to overwritten space in memory.