Security Feed

Improper neutralization of parameters in Snowflake CLI versions prior to 3.19 allowed unintended SQL execution. An attacker could exploit this by supplying crafted values to vulnerable command paths, causing Snowflake CLI to execute unintended SQL in the context of the user’s Snowflake session. Successful exploitation required crafted values to reach vulnerable parameters, including through socially engineered input, malicious repository configuration, or compromised automation feeding external

Gigamon GVOS v5.16.1 and below is vulnerable to Directory Traversal in the GVOS H-VUE subsystem.

U-SPEED AC1200 Gigabit Wi-Fi Router (Model: T18-21K) V1.0 is vulnerable to Command Injection. The Network Time Protocol (NTP) configuration interface does not properly sanitize user-supplied input. An authenticated user with permission to configure NTP settings can inject arbitrary system commands through crafted input fields. These commands are executed with elevated privileges, leading to potential full system compromise.

Mythic before 3.4.0.60 contains an authorization bypass vulnerability in four REST endpoints (c2profile_config_check_webhook, c2profile_redirect_rules_webhook, c2profile_get_ioc_webhook, c2profile_sample_message_webhook) that fail to verify payload ownership. An operator in one operation can invoke these endpoints with a known payload UUID from another operation to access that operation's C2 profile configuration including encryption keys and callback parameters.

Mythic before 3.4.0.60 contains an authorization bypass vulnerability that allows authenticated spectator-role users to perform unauthorized write operations by accessing the eventing_import_automatic_webhook endpoint registered under spectator-permitted middleware. Attackers with spectator role can exploit this misconfigured access control to create and delete automation workflows, making unauthorized modifications to operation automation configuration and EventGroups.

Threat actors are continuing to exploit a critical Langflow vulnerability as part of fresh attacks designed to deliver a Monero cryptocurrency miner. The activity has been found to weaponize CVE-2026-33017 (CVSS score: 9.3), an unauthenticated remote code execution (RCE) vulnerability in Langflow, indicating threat actors are scanning and targeting exposed artificial intelligence (AI)

IBM Db2 on Cloud Pak for Data and Db2 Warehouse on Cloud Pak for Data versions 4.8,5.0,5.1,5.2,5.3 could allow an authenticated user to cause a denial of service when creating new databases due to improper allocation of resources.

IBM Db2 on Cloud Pak for Data and Db2 Warehouse on Cloud Pak for Data versions 4.8, 5.0, 5.1, 5.2, 5.3 could allow a privileged user to perform operations and obtain sensitive information outside of their authority due to improper token validation.

OliveTin is a self-hosted web UI for exposing predefined shell commands to end users. This repository contains a proof-of-concept demonstrating two independent OS command injection vectors in OliveTin's Shell mode execution path, both of which bypass the application's intended shell-argument safety checks.

A malicious extension in the Chrome Web Store is masquerading as the Perplexity AI answer engine, intercepting search traffic and collecting browsing information. [...]

The Tempo and Loki datasource plugins construct backend HTTP requests by interpolating user-supplied input into URL paths without sanitization, enabling path traversal. A Viewer-role user can: (1) capture admin-configured datasource credentials (secureJsonData custom headers) by traversing to an attacker-controlled endpoint, (2) invoke state-changing admin endpoints on Tempo (e.g. /flush, /shutdown), and (3) exfiltrate internal service data via Loki's CallResource which returns full HTTP respons

Cybersecurity researchers have flagged an active browser extension campaign that is designed to steal cryptocurrency by stealthily replacing wallet addresses when unsuspecting users initiate a transaction. The cryptocurrency clipper activity has been codenamed Silent Swap by McAfee Labs. "The campaign is delivered through unsigned installers – observed in both .NET and Golang variants – that

The Snowflake datasource allows for GET/PUT commands, which can allow any user with access to run queries against the data source to read/write files between the local grafana server and the connected Snowflake host.

The geomap panel's XYZ tile layer has a sanitize-then-interpolate ordering bug. sanitizeTextPanelContent() runs on the raw template string before getTemplateSrv().replace() substitutes the variable value, which uses the glob format with no HTML escaping. The result is passed to OpenLayers via element.innerHTML. An Editor can set a textbox variable's default value to an XSS payload that executes for every user who opens the dashboard. This is a bypass of the CVE-2023-0507 fix

The Setracker2 Android Companion App (com.tgelec.setracker) versions 3.1.5 and earlier uses MD5 to generate a request signature for authenticating communications between the mobile client and the backend REST API. Attackers could potentially reverse the signature to recover the session ID. With the session ID exposed, an attacker could impersonate the legitimate user and issue authenticated API requests.

pip would treat console_scripts and gui_scripts as paths instead of file names without sanitizing the resolved absolute path to the installation directory, leading to entry points being installed outside the installation directory.

LLaMA-Factory through 0.9.5 contains a remote code execution vulnerability that allows attackers with WebUI access to execute arbitrary Python code by supplying a malicious model path in the Chat or Training interfaces. The application passes user-supplied model path input unvalidated into AutoTokenizer.from_pretrained() and AutoModel.from_pretrained() with a hardcoded trust_remote_code=True parameter, causing the Hugging Face transformers library to fetch and execute arbitrary code from a remot

Gitea act_runner with the Docker backend (through act 0.262.0) passes a workflow's container.options string to the Docker job container's HostConfig and, when configured with privileged: false, forces only the Privileged flag off while merging options such as --pid=host, --cap-add, and --security-opt unchanged. A user who can run a workflow on a Docker-backed runner can create a job container with host namespaces and broad capabilities and escape to the host as root despite privileged mode being