Researchers have identified a series of vulnerabilities in LangGraph, an open-source framework designed for building AI applications. Among these flaws is a critical SQL injection vulnerability that could allow attackers to execute remote code on affected systems. This is particularly concerning for developers and organizations using LangGraph for self-hosted AI projects, as it could lead to unauthorized access and control over their applications. The vulnerabilities have been patched, but the incident serves as a reminder of the risks associated with using open-source software without proper security measures. Users are advised to update to the latest version to mitigate these risks.
Articles tagged "Vulnerability"
Found 933 articles
CriticalCVE-2026-39987
Attackers Use LLM Agent for Post-Exploitation After Marimo CVE-2026-39987 ExploitThe Hacker News
Actively Exploited
A threat actor has been exploiting a vulnerability in Marimo notebooks, specifically CVE-2026-39987, to gain unauthorized access. After taking control of a publicly accessible notebook, the attacker utilized a large language model (LLM) agent to carry out further actions. They extracted cloud credentials from the compromised system, which could potentially lead to additional breaches or data leaks. This incident raises concerns for organizations using Marimo products, as it demonstrates how quickly attackers can adapt and use advanced tools for post-exploitation activities. Companies must remain vigilant and ensure their systems are secured against such vulnerabilities.
Actively Exploited
Hackers are exploiting a vulnerability in FortiClient Enterprise Management Server (EMS), identified as CVE-2026-35616, which allows them to bypass authentication. This flaw is being used to deliver a credential-stealing malware known as EKZ. Organizations using FortiClient EMS are at risk, as attackers can gain unauthorized access to sensitive information through this exploit. The situation is concerning since the malware targets credentials, potentially leading to further data breaches. Companies should prioritize patching this vulnerability to protect their systems and data from compromise.
Critical
Critical Gogs RCE Vulnerability Lets Any Authenticated User Execute Arbitrary CodeThe Hacker News
A serious vulnerability has been found in Gogs, a widely used open-source Git service that allows users to host their own repositories. This flaw, which has a CVSS score of 9.4, enables any authenticated user to execute arbitrary code, potentially giving them full control over the server. This means that individuals with valid access can exploit this weakness to run malicious commands, posing a significant risk to the integrity and security of the affected systems. Currently, there is no CVE identifier linked to this vulnerability, which may complicate tracking and response efforts. Users of Gogs should be particularly vigilant and consider implementing immediate security measures to mitigate potential exploitation.
CriticalCVE-2026-35616
CVE-2026-35616: FortiClient EMS Flaw Actively Exploited in Malware AttacksSecurity Affairs
Actively Exploited
A recently identified vulnerability in FortiClient Endpoint Management Server (EMS), tracked as CVE-2026-35616, is being actively exploited to deploy information-stealing malware, according to a report from Arctic Wolf. This flaw has a high severity rating of 9.1 and allows attackers to execute remote code without needing authentication, making it particularly dangerous. Organizations using FortiClient EMS should be on high alert as the vulnerability can be exploited through specially crafted requests. The vulnerability was patched in April, but the ongoing exploitation highlights the importance of timely updates and monitoring for suspicious activity. Companies must ensure they have applied the latest patches to protect their systems from these attacks.
Actively Exploited
A newly discovered zero-day vulnerability in the Gogs self-hosted Git service allows attackers to execute remote code on servers that are exposed to the internet. This flaw poses a significant risk to organizations using Gogs for version control, as malicious actors could potentially gain full control over affected systems. Currently, there are no patches available to fix this issue, leaving users vulnerable until a solution is released. The exploitation of this vulnerability is particularly concerning because it can lead to data breaches or further attacks within an organization's infrastructure. Users and administrators of Gogs should take immediate action to secure their installations and monitor for any unusual activity.
Microsoft Condemns "Uncoordinated" Zero Day Disclosures
Infosecurity Magazine
Microsoft has raised concerns about the recent disclosure of several unpatched security vulnerabilities without prior notice. The company stated that these uncoordinated disclosures could put customers at significant risk by exposing them to potential attacks. This situation affects users of Microsoft's products, as they may not be aware of the vulnerabilities or have the necessary patches to protect their systems. The lack of coordinated communication from researchers or security firms can lead to confusion and increased vulnerability. Microsoft urges that such disclosures be handled responsibly to ensure that users are adequately protected and informed.
A recent study by Cisco has revealed that multi-turn prompt injection attacks pose a significant risk to major AI models. These attacks are not effectively measured by success rates from single-turn interactions, which may mislead developers about the safety of their systems. The findings suggest that attackers can manipulate conversations with AI models over multiple exchanges, potentially leading to unintended responses or actions. This vulnerability impacts various AI systems that rely on conversational capabilities, raising concerns about the security of user data and the integrity of AI-generated content. Developers and organizations using these models need to reassess their security measures to protect against these sophisticated attack methods.
A recently discovered security vulnerability in Gitea, a popular self-hosted Git service, has put around 30,000 deployments at risk. The flaw allows attackers to access private container images, which can lead to the exposure of sensitive information such as source code and user credentials. This vulnerability affects organizations that rely on Gitea for their development workflows, making it crucial for them to act swiftly to secure their systems. The situation raises concerns about the security of self-hosted services and the potential for misuse of exposed data. Companies using Gitea should prioritize patching their installations to mitigate this risk.
Researchers from Novee have identified a serious vulnerability in Pretalx, an open-source tool widely used for managing conference talks. This flaw allows attackers to take over accounts, effectively granting them a 100% acceptance rate for talk submissions. This means that malicious users could manipulate the platform to ensure their presentations are accepted without proper vetting. The implications are significant for conference organizers who rely on Pretalx for managing submissions, as it undermines the integrity of the selection process. Users and event organizers are urged to review their security practices and stay alert for updates from Pretalx to protect against potential exploitation.
All Major LLMs Exposed to Multi-Turn Manipulation, Warn Researchers
Infosecurity Magazine
Researchers have discovered that all major large language models (LLMs) are vulnerable to a type of manipulation called multi-turn manipulation. This means that attackers could exploit these models to generate misleading or harmful content over multiple interactions, potentially affecting how users perceive information. The models at risk include those from leading companies in the AI space, which could have serious implications for users relying on these technologies for accurate information. The research highlights the need for developers to implement stronger safeguards against such manipulations, as the integrity of AI-generated content is essential for trust and safety in various applications. This vulnerability raises concerns about the reliability of AI systems, especially when used in sensitive areas like healthcare, finance, and education.
Actively Exploited
As artificial intelligence tools enhance phishing and credential theft techniques, security teams are struggling to keep pace with cybercriminals. The increasing sophistication of these attacks means that stolen credentials are becoming a major vulnerability for organizations. This situation creates a significant risk for companies and their users, as attackers can easily bypass traditional security measures. Organizations must prioritize improving their defenses against credential abuse to protect sensitive data and maintain trust with their customers. The ongoing battle between attackers and defenders highlights the urgent need for more effective security protocols and user education around credential safety.
Researchers have identified a serious vulnerability in Gitea, an open-source platform used for version control, that allows unauthorized users to access private container images. This flaw, labeled CVE-2026-27771, impacts all versions of Gitea prior to 1.26.2. Attackers can exploit this weakness without needing any credentials, which could lead to unauthorized access to sensitive data stored in container images. Given the nature of Gitea as a self-hosted solution, organizations using outdated versions are particularly at risk. It’s crucial for users to update their installations to the latest version to safeguard their private resources.
Actively Exploited
The U.S. Cybersecurity and Infrastructure Security Agency (CISA) has issued an urgent warning to federal agencies, giving them only four days to patch a serious vulnerability in the LiteSpeed cPanel user-end plugin. This flaw is currently being exploited in active attacks, raising significant concerns about the security of servers using this software. Agencies are urged to take immediate action to protect their systems from potential breaches. The situation emphasizes the need for quick responses to known vulnerabilities, especially in government infrastructure, where the impact of a security breach could be severe. Failure to address this could lead to unauthorized access and data compromise.
Critical
Microsoft SharePoint Has a New RCE Flaw. If You Haven’t Patched Yet, Go Do That.Security Affairs
Microsoft has identified a serious vulnerability in SharePoint, labeled CVE-2026-45659, which has a CVSS score of 8.8. This flaw allows attackers to execute remote code with minimal effort, posing a significant risk to organizations using the platform. The vulnerability does not require complicated conditions for exploitation, which increases its potential impact. Microsoft has released security updates to address this issue, and users are strongly advised to apply these patches as soon as possible to protect their systems. Ignoring this vulnerability could lead to unauthorized access and control over affected SharePoint environments.