Stanza is a Stanford NLP Python library for tokenization, sentence segmentation, NER, and parsing of many human languages. Prior to 1.12.2, Stanza model loaders such as stanza.models.common.pretrain.Pretrain.load() attempt torch.load(..., weights_only=True) but fall back to torch.load(..., weights_only=False) on attacker-controllable pickle.UnpicklingError, allowing a malicious .pt pretrain or model file to execute arbitrary pickle code when a Stanza NLP pipeline loads it. This issue is fixed in version 1.12.2.

Wasmtime is a runtime for WebAssembly. Prior to 24.0.11, 36.0.12, 45.0.3, and 46.0.1, wasmtime-wasi hard-link creation and renaming check directory permissions but not matching FilePerms on source and destination preopens, allowing a WASI guest with a read-only source file capability to overwrite host files exposed as FilePerms::READ through wasip1, wasip2, or wasip3 filesystem interfaces. This issue is fixed in versions 24.0.11, 36.0.12, 45.0.3, and 46.0.1.

UMTS FP protocol dissector crash in Wireshark 4.6.0 to 4.6.6 and 4.4.0 to 4.4.16 allows denial of service

LiteLLM is a proxy server (AI Gateway) to call LLM APIs in OpenAI (or native) format. Prior to 1.82.0-stable, LiteLLM's Custom Code Guardrails production create and update paths did not apply the same sandboxing and validation used by the test endpoint, allowing a privileged user with access to create or update guardrails to submit custom Python code that executed in the LiteLLM proxy environment and could expose secrets available to the process. This issue is fixed in version 1.82.0-stable.

Integer overflow in Extensions API in Google Chrome prior to 150.0.7871.115 allowed an attacker who convinced a user to install a malicious extension to perform an out of bounds memory read via a crafted Chrome Extension. (Chromium security severity: High)

n8n before 1.123.61, 2.x before 2.27.4, and 2.28.x before 2.28.1 contains a SQL injection vulnerability in the legacy MySQL v1 node's executeQuery operation. The operation substitutes evaluated {{ ... }} expression values directly into the raw SQL string without parameterization. When a workflow uses this operation with expression-sourced values and is connected to an externally-reachable trigger (such as a Webhook node), attacker-controlled input reaching those expressions results in SQL injection, allowing execution of arbitrary SQL with the configured MySQL credentials' privileges. The MySQL v2 node, which uses parameterized queries, is not affected.

n8n before versions 1.123.18 and 2.6.2 fails to verify HMAC-SHA256 signatures on Zendesk webhooks in the ZendeskTrigger node. Attackers who know the webhook URL can send unsigned POST requests to trigger workflows with arbitrary malicious data.

A vulnerability was found in bentoml OpenLLM 0.6.30. This affects the function async_run_command of the file src/openllm/common.py of the component Model Repository Directory Name Handler. Performing a manipulation of the argument cmd results in command injection. Attacking locally is a requirement. The exploit has been made public and could be used. The project was informed of the problem early through an issue report but has not responded yet.

Insufficient validation of untrusted input in Codecs in Google Chrome on Windows prior to 150.0.7871.115 allowed a remote attacker who had compromised the renderer process to potentially perform a sandbox escape via a crafted HTML page. (Chromium security severity: High)

Composer is a dependency Manager for the PHP language. Prior to 2.2.29 and 2.10.2, a maliciously crafted package from an untrusted repository other than Packagist.org or Private Packagist can cause Composer to write attacker-controlled files outside the vendor directory and outside the project during install or update by using an invalid package name that is not correctly validated before dependency-resolution results are written or installed. This issue is fixed in versions 2.2.29 and 2.10.2.