Overview of incoming advisories.
2179 results found
Security researchers at Parameter discovered a critical vulnerability in ChatGPT's shared link feature that allowed attackers to execute code within a victim's active session. Any logged-in user who opened a malicious ChatGPT share link was at risk of having their private conversations read by the attacker. The flaw was reported to OpenAI via their bug bounty program on October 18 and was patched eight days later. OpenAI classified the vulnerability as P1, their highest severity rating. The attack required no special privileges from the victim beyond simply clicking a shared link while logged in.
View original advisory →On October 29, 2026, a malicious version (2.1.313) of the Claude Code VS Code extension was published to Cursor's extension marketplace. The trojanized extension reached up to 5,000 Cursor users during an 11-hour exposure window. Upon loading, the extension planted an attacker-controlled SSH key on the victim machine and established persistence mechanisms tailored to each operating system. It also opened a reverse SSH tunnel back to attacker-controlled infrastructure, enabling persistent remote access. Additionally, the malware collected host reconnaissance data and transmitted it over unencrypted plain HTTP. The attack represents a supply chain compromise targeting software developers using the Cursor IDE environment. The incident highlights risks associated with third-party extension marketplaces and the potential for developer machine compromise via IDE plugins.
View original advisory →Wiz research documents a wave of agentic attacks targeting AI developer platforms during October, dubbed 'Black October.' Attackers are exploiting vulnerabilities and misconfigurations in AI development environments to perform LLMJacking at scale. LLMJacking involves hijacking access to large language model APIs and compute resources for unauthorized use. The attacks leverage agentic AI capabilities to automate and amplify the scope of compromise. This represents an emerging and growing threat vector specifically targeting the AI/ML development ecosystem. The incidents highlight the need for stronger security controls around AI platform credentials, API keys, and access management.
View original advisory →Security researchers at Anthropic have discovered and developed an exploit for a critical vulnerability in the Linux kernel affecting servers running Multipath TCP (MPTCP) listeners. The flaw allows unauthenticated remote attackers to execute arbitrary commands with root privileges, making it a severe remote code execution vulnerability. No authentication is required to exploit the issue, significantly raising its risk profile. The vulnerability targets MPTCP, a transport protocol extension that allows multiple network paths to be used simultaneously. Servers exposed to the internet running MPTCP listeners are at high risk of full compromise. The research highlights the danger of kernel-level vulnerabilities that can be exploited remotely without credentials. System administrators running Linux servers with MPTCP enabled should apply patches or mitigations immediately.
View original advisory →TeamViewer has patched a security vulnerability that allowed a controlled machine (host) to attack the controlling machine (viewer) during a remote session. In a typical TeamViewer session, the viewer connects to and controls the host. This flaw reversed the attack vector, enabling a malicious or compromised host to crash the viewer's TeamViewer client. Depending on the platform, the vulnerability may also allow arbitrary code execution on the viewer's machine. The flaw represents a significant risk as it subverts the expected trust model of remote desktop sessions. Users who connect to untrusted or compromised machines could be at risk. TeamViewer has issued a fix and users are advised to update immediately.
View original advisory →A malicious version of Anthropic's Claude Code VS Code extension was distributed via Cursor's extension marketplace for approximately 11 hours on October 29. The fake extension exposed up to 5,000 developers to malware designed to install persistent SSH backdoors on their systems. The attack targeted software developers using the Cursor IDE, a popular AI-assisted coding environment. The malicious package masqueraded as a legitimate tool to gain trust and installation. SSH backdoors allow attackers persistent remote access to compromised machines. The incident highlights risks in third-party extension marketplaces for developer tools. The supply chain attack vector is particularly concerning given the high-privilege environments developers typically operate in.
View original advisory →A high-severity remote code execution vulnerability has been discovered in WordPress Core affecting versions 4.7 through 7.1.3, carrying a CVSS score of 8.1. The flaw does not yet have a CVE identifier assigned. Patches were released and almost immediately threat actors began attempting to exploit the vulnerability. Within just 24 hours of patch release, nearly 63,000 attack attempts were observed targeting this flaw. The rapid exploitation attempts highlight the urgency for WordPress site administrators to apply available patches without delay. The vulnerability allows attackers to execute arbitrary code remotely on affected WordPress installations.
View original advisory →CISA has warned that a maximum-severity GitLab vulnerability is now being actively exploited in the wild. The flaw was patched earlier in the month but attackers have begun leveraging it in real-world attacks. CISA's alert signals that organizations using GitLab should prioritize patching immediately. The vulnerability carries the highest severity rating, indicating significant potential impact. The exploitation has prompted a formal warning from the US cybersecurity agency, underscoring the urgency of remediation for affected GitLab installations.
View original advisory →Drupal has issued a warning to administrators about a critical access bypass vulnerability in the Commerce Core module. The flaw is being actively exploited in the wild, making it a high-priority security concern. Drupal is urging all affected site owners to apply the available patch immediately. The vulnerability could allow attackers to bypass access controls within the Commerce Core functionality. Active exploitation in real-world attacks elevates the urgency of remediation. No further details about specific CVE identifiers or threat actors were mentioned in the available content.
View original advisory →On October 29, 2026, version 2.1.313 of the official Claude Code VS Code extension (anthropic.claude-code) was published with a malicious loader embedded in its extension.js bundle. The trojanized version was distributed via the Open-VSX registry and Cursor's extension marketplace, reaching up to 5,000 Cursor users during an 11-hour window. Upon loading, the extension planted an attacker-controlled SSH key on the victim machine, installed OS-specific persistence mechanisms, and opened a reverse SSH tunnel to attacker-controlled infrastructure. Additionally, the malware beaconed host reconnaissance data over plain HTTP. The attack represents a supply chain compromise targeting developer machines through a trusted IDE extension. StepSecurity verified the payload behavior through sandbox and independent registry analysis. The incident highlights risks in mirrored extension marketplaces where malicious versions can propagate quickly.
View original advisory →Cursor confirmed a malicious Claude Code extension version 2.1.313 was available through its marketplace on October 29, potentially affecting up to 5,000 users. The extension contained code to add SSH authorized keys, establish a reverse SSH tunnel, and install persistence mechanisms. Socket's investigation found no evidence that this version was published to Open VSX or npm registries, challenging initial reports attributing distribution to those channels. The original report by OpenSourceMalware has since been revised, removing Open VSX as the identified distribution channel. Screenshot evidence in the original report showed suspicious discrepancies, with file sizes matching version 2.1.314 rather than 2.1.313. Cursor responded by removing the extension, rebuilding its marketplace with rotated keys, and working with hosting providers to take down the attacker's server. Cursor published remediation guidance including package hashes and indicators of compromise for the malicious macOS and Windows packages.
View original advisory →Binance, the world's largest cryptocurrency exchange, has attributed a massive theft of approximately $6.3 billion in cryptocurrency to a compromised laptop belonging to a DevOps engineer. The incident highlights the risks posed by insider endpoint compromise in high-value financial environments. A single engineer's compromised workstation served as the entry point for attackers to access and steal funds at a massive scale. This case underscores the critical importance of endpoint security, privilege management, and access controls in cryptocurrency and financial organizations. The breach represents one of the largest crypto thefts linked to a single compromised device, raising significant concerns about DevOps security practices and the protection of privileged access in the crypto industry.
View original advisory →The official Anthropic Claude Code extension on the Open VSX registry was compromised to deliver a persistent SSH backdoor on infected systems. The attack was traced back to a compromise of the Cursor Marketplace, an AI coding platform, which temporarily took down its service to remove the malware. The backdoor was designed to maintain persistent access to victim machines via SSH. This represents a significant supply chain attack targeting developer tools and AI coding assistants. The incident highlights risks associated with third-party extension marketplaces used by developers. Cursor took immediate remediation steps upon discovery of the malicious code.
View original advisory →A highly critical security vulnerability tracked as CVE-2026-118539 has been discovered in Drupal's Commerce Core component. The flaw is a missing authorization issue affecting Commerce Core versions 0.0.0 to 2.41 and 3.0.0 to 3.3.13. Threat actors are actively exploiting this vulnerability in the wild. The Commerce Core module enables e-commerce functionality including shopping cart capabilities for Drupal-based websites. Drupal has issued a warning to its user base regarding the active exploitation. The CVSS score has not yet been assigned, though the vulnerability is classified as highly critical. Site administrators are urged to apply patches or mitigations immediately to prevent unauthorized access.
View original advisory →Binance, the world's largest cryptocurrency exchange, disclosed a major security incident affecting approximately $6.3 billion worth of assets. Nearly $3 billion of the impacted assets have been frozen on-chain as part of the response. Stablecoin issuers Tether and Circle took action, freezing over 1.567 billion USDT and 800 million USDT respectively linked to the incident. The breach represents one of the largest security incidents in cryptocurrency exchange history. The full scope of the attack vector and attribution has not been disclosed in the available content. The freezing of assets by stablecoin issuers indicates coordinated industry response to limit further damage. This incident highlights ongoing vulnerabilities in centralized cryptocurrency exchange infrastructure.
View original advisory →CVE-2026-111357 is a missing authentication for critical function vulnerability in Ivanti Connect Secure that allows a local authenticated attacker to escalate privileges. The vulnerability can be chained with two other vulnerabilities, CVE-2026-104935 and CVE-2026-118527, potentially increasing the severity of exploitation. CISA has issued an alert urging immediate patching and has referenced BOD 26-04 which prioritizes security updates based on risk. Forensic triage requirements have also been outlined by CISA for affected organizations. The vulnerability is actively exploited and tracked on the NVD. Ivanti has published a security advisory with remediation guidance. Organizations using Ivanti Connect Secure are strongly urged to apply available patches without delay.
View original advisory →Ivanti Connect Secure contains a vulnerability involving the generation of predictable numbers or identifiers (CVE-2026-118527), allowing a remote authenticated attacker to upload files with attacker-controlled content to a known location on disk. This vulnerability is notable for its chaining potential with two additional CVEs: CVE-2026-104935 and CVE-2026-111357, increasing overall exploitability and impact. The vulnerability is tracked by CISA and is subject to BOD 26-04, which mandates prioritizing security updates based on risk. Forensic triage requirements have also been outlined by CISA for affected organizations. The issue is documented on the NVD and has an official security advisory from Ivanti. Organizations using Ivanti Connect Secure are urged to apply patches and follow CISA guidance immediately. The current risk level is rated High.
View original advisory →Cryptocurrency exchange Binance is investigating a potential security breach that could involve up to $5.8 billion in digital assets. If confirmed, this could become the largest cryptocurrency hack in history. The investigation is ongoing and details remain limited. The incident raises significant concerns about the security of major cryptocurrency exchanges and the safety of user funds. The scale of the potential theft underscores the high-value targets that crypto platforms represent for threat actors.
View original advisory →A crypto-mining campaign targeting Hugging Face models has returned shortly after being taken down, demonstrating persistence and resilience. The attackers are hijacking GPU servers by embedding malicious code within Hugging Face model files. One malicious model was downloaded over 28,000 times before detection, indicating significant reach. The campaign also includes planting backdoors on compromised systems for persistent access. In a related but separate incident, a fake Solana starter kit was discovered on GitHub designed to trick AI coding agents into installing NetSupport RAT. The dual-pronged attack highlights the growing abuse of AI/ML platforms and developer tools as attack vectors. The return of the campaign after takedown underscores the challenges of content moderation on large model-sharing platforms. GPU hijacking for crypto mining represents a significant financial threat to cloud and AI infrastructure operators.
View original advisory →Polymarket, a prediction market platform, temporarily halted trading after hackers compromised a third-party service to inject cryptocurrency-draining JavaScript into the platform. Approximately $2.5 million in assets was traced to a suspected attacker wallet. The attack involved a supply chain-style compromise where malicious JavaScript was distributed through a third-party provider. The incident highlights risks associated with third-party script dependencies in financial platforms handling cryptocurrency. Polymarket suspended trading as a precautionary measure while investigating the breach. The attack is consistent with a growing trend of crypto-draining scripts targeting high-value DeFi and prediction market platforms.
View original advisory →