CVE-2026-70085 is a medium severity vulnerability in Microsoft Windows Server 2012 DNS Server involving missing authorization. This flaw allows an authorized local attacker to tamper with DNS Server components without proper authorization controls. The issue affects multiple Windows Server 2012 and related versions and has been officially fixed by Microsoft.
CVE-2026-75120 is a critical authentication bypass vulnerability in the OpenText™ Endpoint Investigator product family. It affects the EnCase SAFE Agents component used by several OpenText security products. The flaw allows unauthorized or unauthenticated actors to establish unauthorized connections to vulnerable Agents by bypassing security controls during session establishment.
PickMall lilishop versions up to 4.4 contain an insecure direct object reference vulnerability that allows authenticated sellers to modify other stores' product freight templates. This occurs because the update functions for goods and SKUs do not enforce store-scoped access controls, enabling sellers to alter shipping charges on competitors' products.
Meta experienced a security breach where hackers accessed some of its GitHub repositories using stolen employee credentials. Approximately 1,500 individuals, mostly Meta employees in the United States, had their personal data and business-related software code exposed. The personal data was mistakenly stored in the repositories against company policy, and much of it was synthetic or anonymized test data. Meta's internal systems were not breached, and the incident was not due to a vulnerability in Meta products. The company has taken steps to rotate secrets, enhance security controls, and prevent user data from being stored on GitHub in the future.
CVE-2026-95680 is a high-severity vulnerability in IBM WebSphere Application Server versions 8.5 and 9.0. It involves improper authentication (CWE-287) that could allow an authenticated attacker to bypass authorization controls, potentially leading to administrative impersonation across the server cell. This could result in full compromise of administrative functions. The vulnerability has a CVSS 3.1 score of 8.0, indicating a significant risk. No vendor patch or remediation details are provided in the available data.
CVE-2026-74033 is a path traversal vulnerability in Wazuh wazuh-manager versions from 4.1.1 up to but not including 4.14.8. It affects the internal cluster RPC path handlers get_lists() and delete_list_file(), allowing compromised cluster workers to bypass REST validation and RBAC controls. This enables attackers to read or delete sensitive master files such as rbac.db. The vulnerability is fixed in version 4.14.8.
CVE-2026-109116 is an authorization bypass vulnerability in the AdithyaYelloju Restaurant-Management-System affecting the Order Management component. The flaw arises from manipulation of the ID argument in the mysqli_query function, allowing remote attackers to bypass authorization controls. The product does not use versioning, so affected versions cannot be precisely identified. The vulnerability has a medium severity with a CVSS score of 6.9. The issue was reported early to the project, but no response or patch has been provided yet. Exploit details have been publicly disclosed, though no active exploitation in the wild is currently known.
CVE-2026-13138 is a high-severity vulnerability in the Biletiva Ticketing Platform by Divasoft Software and E-Commerce Inc. It involves client-side enforcement of security controls that should be handled server-side, allowing manipulation of application state. The vulnerability affects versions up to and including 2026-07-23. It has a CVSS 3.1 score of 8.6, indicating a significant risk. No patch or remediation details are provided, and no known exploits are reported in the wild.
A critical security vulnerability was discovered in a SmartOffice biometric and HRMS system instance used by Hyderabad-based companies. The vulnerability stemmed from a weak password policy that allowed an attacker to gain full administrative access, exposing sensitive employee data, login records, biometric device controls, and HR functions. The issue was responsibly disclosed to the vendor and affected organizations. This highlights the importance of strong password policies and proper network security configurations.
Abu Dhabi is implementing AI-agent security controls across more than 50 government entities to mitigate risks where attackers might manipulate AI agents into performing unauthorized actions that appear legitimate. The main concern is not AI errors but malicious instructions causing agents to access sensitive data or execute unauthorized transactions. The security strategy involves defense in depth with monitoring, auditing, guardrails, and kill switches to shut down misbehaving AI agents. Experts emphasize the need for real-time monitoring and clear identities for AI agents to prevent them from becoming malware-like threats. The approach aims to build trust in AI-native government operations while addressing emerging attack techniques such as 'living off local agents' and malicious AI skills.
CVE-2026-108924 is an authorization bypass vulnerability in the getkirby Kirby CMS affecting versions 4.9.0 through 4.9.5 and 5.5.0 through 5.5.3. The issue exists in the OptionsQuery::resolve function within the Dynamic Option Generation component. This vulnerability allows remote attackers to bypass authorization controls. Upgrading to versions 4.9.6 or 5.6.0 resolves the issue.
CVE-2026-108923 is a medium severity vulnerability in 1024-lab smart-admin up to version 3.30.0. It involves improper access controls in the /support/config/queryByKey functionality of the Support Config Query component. The vulnerability arises from manipulation of the configKey argument, allowing remote attackers to bypass access restrictions. The issue was reported early to the project, but no response or patch has been provided yet. Exploit details have been publicly disclosed, though no known exploits are currently observed in the wild.
CVE-2026-87744 is a critical vulnerability in the Red Hat build of Quarkus 3.27.6.redhat affecting the REST Data Panache component. It allows unauthenticated attackers to bypass security controls on generated REST endpoints when developers customize operations using @MethodProperties without re-declaring security roles. This causes the code generator to drop class-level security annotations, resulting in endpoints defaulting to permissive, unauthenticated access. This can lead to unauthorized information disclosure or data modification.
CVE-2026-108796 is an authorization bypass vulnerability in ZKTeco BioTime version 9.5.6 Build 20260130.3206 affecting the Biometric Data Endpoint component. The flaw involves manipulation of the emp_code argument in the /iclock/finger_data/ file, allowing remote attackers to bypass authorization controls. The issue is mitigated by upgrading to version 9.5.6 Build 3623.
Sylius versions through 2.3.0 have an authorization bypass vulnerability that allows unauthenticated attackers to access unmoderated and rejected product reviews. This occurs because the AcceptedExtension filter is not applied to the item operation, enabling attackers to enumerate sequential review IDs via the API endpoint /api/v2/shop/product-reviews/{id}. Attackers can retrieve review titles, ratings, comments, timestamps, and author first names, bypassing merchant moderation controls.
CVE-2026-108575 is a medium severity vulnerability in BerriAI LiteLLM up to version 1.94.0. It involves improper authorization in the get_secret function of the Secret Resolution component, specifically in secret_managers/main.py. The vulnerability arises from manipulation of the api_key argument, allowing remote attackers to bypass authorization controls. The issue has been publicly disclosed, but the vendor has not responded or provided a fix.
CVE-2026-106609 is a high-severity vulnerability in the Bayarcash WooCommerce plugin by Web Impian. It involves missing authorization controls, allowing unauthorized users to exploit incorrectly configured access control security levels. This affects versions from initial releases up to 4.4.2 inclusive. The vulnerability does not impact confidentiality but allows unauthorized modification of data, with no impact on availability. No known exploits are reported in the wild, and no official patch information is provided.
This report discusses the security risks posed by AI agents using valid credentials to perform actions beyond their assigned permissions, particularly in corporate environments. It highlights how traditional access controls may fail to prevent agents from escalating privileges by switching to more powerful credentials available on the same system. The article outlines various enforcement points and mitigation strategies to limit agent actions without sacrificing autonomy, including managed agent settings, runtime hooks, gateways, sandboxes, endpoint enforcement, and credential authorization. It emphasizes the importance of scoping agent permissions tightly and enforcing controls at the tool call level to prevent unauthorized actions. The report does not describe a specific vulnerability or exploit but rather addresses a broader security challenge in managing AI agent permissions.
CVE-2026-8374 is a vulnerability in the SwitchBot Lock series caused by misuse and misconfiguration of Bluetooth communication. This flaw allows an attacker to bypass the electronic lock and access controls by manipulating the communication protocol. The issue stems from the generation of weak initialization vectors (IVs) in the Bluetooth communication process. Multiple versions of the SwitchBot Lock firmware prior to fixed releases are affected. The vulnerability has a high severity rating with a CVSS score of 8.5.
CVE-2026-17106 is a vulnerability in the moby/go-archive component used by Red Hat Quay. The flaw allows an attacker who controls the contents of a tar archive to exploit symbolic link handling during extraction, enabling arbitrary file writes outside the intended directory. This can lead to unauthorized modification of system files or potentially arbitrary code execution. The vulnerability is rated as important/high severity by Red Hat. Mitigation involves avoiding processing tar archives from untrusted sources and running extraction with least privilege. Red Hat Quay 3.12.23 includes this advisory but does not explicitly state a fix for this CVE in the provided content.