A heap-based buffer overflow vulnerability exists in the Sharkd utility of Wireshark versions 4.4.0 through 4.4.18 and 4.6.0 through 4.6.8. This flaw can cause the utility to crash, resulting in a denial of service condition. The vulnerability requires local access with user interaction to be exploited and does not impact confidentiality or integrity.
CVE-2026-95387 is a heap-based buffer overflow vulnerability in the SPDY protocol dissector of Wireshark versions 4.4.0 through 4.4.18 and 4.6.0 through 4.6.8. This vulnerability can cause Wireshark to crash, resulting in a denial of service condition. The flaw has a high severity with a CVSS score of 8.1. No known exploits are reported in the wild. No patch or official fix information is provided in the available data.
CVE-2026-95386 is a medium severity vulnerability in Wireshark versions 4.6.0 through 4.6.8. It involves an infinite loop in the TTL file parser due to a loop with an unreachable exit condition, which can cause a denial of service (DoS) condition. The vulnerability does not impact confidentiality or integrity but affects availability by potentially causing the application to hang or become unresponsive.
Ghostscript for Windows is vulnerable to local privilege escalation through PostScript resource file hijacking. Due to the application searching for PostScript resource files in predictable paths under C:\\gs\\ that do not exist by default on Windows installations, combined with Windows default ACLs allowing any authenticated user to create directories at the root of C:\\, an attacker who is an authenticated local user can create the expected directory structure and plant a malicious PostScript file. When any user or service subsequently runs Ghostscript, the planted file is automatically loaded and executed with the full privileges of the Ghostscript process. This results in full compromise of Ghostscript process context, as well as running arbitrary code on the machine with Ghostscript process privileges.
This issue was fixed in version 10.08.0.
Das U-Boot with CONFIG_IP_DEFRAG=y parameter fails to clear IP reassembly state after delivering a complete datagram. An attacker who can deliver fragmented IP traffic can execute arbitrary code by sending duplicated last-fragment IP packets.
This issue was fixed in commit b1aec609bb5e0d08c25c888c91935287ab4ee5fa in version 2026.07.
OperTraitor is an open-source tool designed to audit Kubernetes operators for excessive RBAC privileges, identifying risks from overly permissive service accounts. Kubernetes operators automate cluster management but often run with broad permissions, creating security weak spots that can be exploited if compromised. The tool compares an operator's documented functionality against its actual granted privileges, assigning a risk score to highlight potential threats. Research using OperTraitor uncovered a high-severity vulnerability (CVE-2026-6389) in IBM's Turbonomic platform and widespread issues with abandoned or overly permissive operators in OperatorHub. The rise of AI-driven agentic operators increases the risk, as autonomous operators with excessive permissions can cause greater damage. The analysis emphasizes securing service accounts and auditing operator permissions before deployment to mitigate these risks.
Executive Threat Detection (ETD) is a proactive cybersecurity service launched by Cisco Talos Incident Response to protect high-value executive accounts from sophisticated, targeted attacks. It provides monthly human-led threat hunting and intelligence analysis focused on executives' corporate systems, addressing the gap left by enterprise-wide security tools tuned for average users. The service integrates open-source intelligence (OSINT) to identify emerging threats such as phishing kits and zero-day exploits targeting executives. ETD hunts for subtle, low-and-slow attack techniques and monitors corporate information for signs of compromise. It operates with minimal impact on executive productivity by leveraging existing security tools without requiring system changes. Monthly reports provide technical findings and strategic recommendations, enabling informed leadership decisions. ETD is designed to complement existing Talos IR retainer services, allowing rapid escalation to emergency response if needed.
The Apache Airflow Teradata provider's compute-cluster example Dag declared every one of its Dag Params as unconstrained free text and templated them straight into the compute-cluster operators, which interpolate those values into Teradata DDL. A user who is permitted to trigger that Dag - a lower-trust role than the Dag author, and one that needs no Teradata credentials of its own - could therefore supply SQL fragments that execute under the connection the task runs as, and could additionally redirect the task at any other connection defined in the deployment, because the connection id was itself a free-text Param. Only deployments that run this example Dag, or a Dag copied from it, are affected; the provider's operator code is unchanged. Users of apache-airflow-providers-teradata are recommended to upgrade to version 3.7.0 or later, whose example constrains the Params to validated identifiers and a closed value set and removes connection selection and free-form option strings from trigger-time input. Upgrading does not change a Dag already copied from the example; users who copied it should apply the same constraints to their copy.
Apache Airflow's Snowflake provider did not validate the connection's `account` and `region` fields before interpolating them into request URLs. The SQL API endpoint is built as `https://{account}.snowflakecomputing.com/api/v2/statements`, so an `account` value containing `/`, `?` or `#` demotes the intended domain to a path, query or fragment and leaves the attacker in control of the request host.
The provider sends that request with an `Authorization: Bearer` header carrying a JWT minted from the connection's private key, or the configured OAuth or programmatic access token. A user who can edit the Snowflake connection but cannot read its secrets — Airflow gives connection-configuration users write-only access to stored credentials, and a `private_key_file` lives on the worker rather than in the connection — can therefore cause a valid token for the account to be delivered to a host of their choosing and replay it against the genuine Snowflake endpoint. No Dag-authoring ability is required: the attacker edits the connection and waits for an existing Dag to use it. The same unvalidated value was also used to build the OAuth token-request URL and the Cortex Agent base URL.
Affects deployments where Snowflake connections are editable by users who are not trusted with the connection's credentials. Users are advised to upgrade to `apache-airflow-providers-snowflake` `6.18.0` or later, which rejects `account` and `region` values containing anything other than letters, digits, `.`, `_` and `-` in every URL the provider builds from them.
Apache Airflow's Teradata provider versions before 3.7.0 embed cloud storage credentials directly into SQL statements, which are then logged in Airflow task logs and Teradata's query logs. This exposure occurs when using the S3ToTeradataOperator or AzureBlobStorageToTeradataOperator against private buckets without a configured Teradata authorization object. The credentials appear in plaintext in Airflow logs and Teradata's DBQL query logs, posing a risk of credential leakage. Upgrading to version 3.7.0 or later prevents credentials from appearing in Airflow logs but does not remove them from Teradata's logs. Users should configure a Teradata authorization object and rotate exposed credentials.
CVE-2026-81914 is a medium severity vulnerability in the Apache Airflow Google provider where Google Drive search queries improperly handle special characters in file and folder names. Specifically, names containing apostrophes are not escaped, allowing an attacker controlling these names to inject additional clauses into the query. This can cause the system to select attacker-controlled files or folders during upload or download operations, potentially redirecting data flows. The issue affects versions prior to 22.6.0. Upgrading to version 22.6.0 or later mitigates the vulnerability by properly escaping special characters in queries.
Asset Suite allows unauthenticated users to access PropertiesReloadServlet, CacheFlushServlet, MetadataCacheFlushServlet and ResourceBundleReloadServlet, which could result in denial-of-service conditions affecting application availability. These servlets are designed to perform specific functions within production environment depending on how the Asset Suite application is configured.
An empirical study evaluated the network exposure of Ray control-plane components in distributed large language model (LLM) inference deployments on Kubernetes using Amazon EKS 1.35. The study found that multiple Ray control-plane sockets were not declared in pod metadata, allowing an unprivileged pod in a different namespace to access Ray GCS and raylet RPC services over unauthenticated cleartext gRPC. Default security scanners did not detect this exposure. A default-deny NetworkPolicy blocked access in steady state, and WireGuard encryption added modest overhead. The Ray Job API responded to unauthenticated requests on CPU Ray images but was not active in GPU deployments. The findings reflect vendor default configurations rather than new vulnerabilities, and a related CVE affecting vLLM was already fixed upstream.
NeedyMantis is a modular malware framework used by a China-based threat actor in targeted attacks against telecommunications, government, scientific, manufacturing, retail, medical non-profit, and intergovernmental organizations primarily in Belarus, Russia, and Thailand. It was discovered following the May 2026 Daemon Tools supply chain attack, where thousands of computers were infected. The malware uses a custom executable file format and DLL sideloading for persistence and evasion. It is deployed after initial access to maintain long-term presence and support follow-on operations. The malware communicates with command-and-control servers via WebSockets and can load additional modules to extend its capabilities. The threat actor group behind this malware, Storm-3069, is not attributed to a Chinese nation-state actor. Microsoft has provided a detailed technical analysis but no patch or remediation guidance is indicated.
Asset Suite allows unauthenticated users to access HTTPPublishAdapterTestServlet that can be used for configuration file upload, leading to information disclosure and integrity compromise. The HTTPPublishAdapterTestServlet is specifically meant for testing purposes to be used in a non-production environment.
NVIDIA OpenShell is an open-source sandbox for AI agents designed to block unauthorized network, file, and process access. In tests of version 0.1.2, the default policy effectively prevented data exfiltration, blocking all attempted leaks. However, four specific operator-configurable settings can allow data to leave the sandbox: read-write access rules, GET-only rules that still transmit data in query strings and headers, audit mode rules that log but do not block, and automatic approval of new public hosts without human intervention. The sandbox does not validate certain rule types (GraphQL, MCP, WebSocket, JSON-RPC), accepting them without warning. The sandbox is effective under default settings but requires careful policy management to maintain security.
An improper authorization vulnerability in the end-of-life versions of RTU500’s web application allows an authenticated user to trigger the RTU500 to reboot through the reset endpoint. Successful exploitation could cause temporary device unavailability and disruption of its intended operation.
