#git

New research shows Google Cloud and smaller providers have the highest cloud vulnerability rates as compared to AWS…

Security policies like [`allowed-gadgets`](https://inspektor-gadget.io/docs/latest/reference/restricting-gadgets), [`disallow-pulling`](https://inspektor-gadget.io/docs/latest/reference/disallow-pulling), [`verify-image`](https://inspektor-gadget.io/docs/latest/reference/verify-assets#verify-image-based-gadgets) can be bypassed by a malicious client. ### Impact Users running `ig` in daemon mode or IG on Kubernetes that rely on any of the features mentioned above are vulnerable to this issue. In order to exploit this, the client needs access to the server, like the correct TLS certificates on the `ig daemon` case or access to the cluster in the Kubernetes case. ### Patches The issue has been fixed in v0.40.0 ### Workarounds There is not known workaround to fix it.

The communications app TeleMessage, which was spotted on former US national security adviser Mike Waltz's phone, has suspended “all services” as it investigates reports of at least one breach.

### Summary A LLM application leveraging `XMLToolMessage` class may be exposed to untrusted XML input that could result in DoS and/or exposing local files with sensitive information. ### Details `XMLToolMessage` uses `lxml` without safeguards: https://github.com/langroid/langroid/blob/df6227e6c079ec22bb2768498423148d6685acff/langroid/agent/xml_tool_message.py#L51-L52 `lxml` is vulnerable to quadratic blowup attacks and processes external entity declarations for local files by default. Check here: https://pypi.org/project/defusedxml/#python-xml-libraries ### PoC A typical Quadratic blowup XML payload looks like this: ```xml <!DOCTYPE bomb [ <!ENTITY a "aaaaaaaaaa"> <!ENTITY b "&a;&a;&a;&a;&a;&a;&a;&a;&a;&a;"> <!ENTITY c "&b;&b;&b;&b;&b;&b;&b;&b;&b;&b;"> ]> <bomb>&c;</bomb> ``` Here, &a; expands to 10 characters, &b; expands to 100, and &c; expands to 1000, causing exponential memory usage and potentially crashing the application. ### Fix Langroid 0.53.4 initializes `XMLParser` with...

The fix to https://cantina.xyz/code/c486d600-bed0-4fc6-aed1-de759fd29fa2/findings/21 has a typo that still results in the highest limb of `pc` being range checked to 8-bits instead of 6-bits. In the AIR, we do https://github.com/openvm-org/openvm/blob/0f94c8a3dfa7536c1231465d1bdee5fc607a5993/extensions/rv32im/circuit/src/auipc/core.rs#L135 ``` for (i, limb) in pc_limbs.iter().skip(1).enumerate() { if i == pc_limbs.len() - 1 { ``` It should be ``` for (i, limb) in pc_limbs.iter().enumerate().skip(1) { ``` Right now the if statement is never triggered because the enumeration gives `i=0,1,2` when we instead want `i=1,2,3`. What this means is that `pc_limbs[3]` is range checked to 8-bits instead of 6-bits. This leads to a vulnerability where the `pc_limbs` decomposition differs from the true `pc`, which means a malicious prover can make the destination register take a different value than the AUIPC instruction dictates, by making the decomposition overflow t...

Craft CMS contains a potential remote code execution vulnerability via Twig SSTI. You must have administrator access and `ALLOW_ADMIN_CHANGES` must be enabled for this to work. https://craftcms.com/knowledge-base/securing-craft#set-allowAdminChanges-to-false-in-production Note: This is a follow-up to https://github.com/craftcms/cms/security/advisories/GHSA-f3cw-hg6r-chfv Users should update to the patched versions (4.14.13 and 5.6.15) to mitigate the issue. ### References https://github.com/craftcms/cms/pull/17026

### Summary During a manual source code review, [**ARIMLABS.AI**](https://arimlabs.ai) researchers identified that the `browser_use` module includes an embedded whitelist functionality to restrict URLs that can be visited. This restriction is enforced during agent initialization. However, it was discovered that these measures can be bypassed, leading to severe security implications. ### Details **File:** `browser_use/browser/context.py` The `BrowserContextConfig` class defines an `allowed_domains` list, which is intended to limit accessible domains. This list is checked in the `_is_url_allowed()` method before navigation: ```python @dataclass class BrowserContextConfig: """ [STRIPPED] """ cookies_file: str | None = None minimum_wait_page_load_time: float = 0.5 wait_for_network_idle_page_load_time: float = 1 maximum_wait_page_load_time: float = 5 wait_between_actions: float = 1 disable_security: bool = True browser_window_size: Browse...

ESET has discovered Spellbinder, a new tool used by the China-linked cyber espionage group TheWizards to conduct AitM…

This week on the Lock and Code podcast, we speak with Emanuel Maiberg and Jason Koebler about Overwatch, an AI chatbot tool sold to US police.

**Vulnerable MobSF Versions:** <= v4.3.2 **CVSS V4.0 Score:** 8.6 (CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:P/VC:H/VI:H/VA:L/SC:N/SI:N/SA:N) **Details:** A Stored Cross-Site Scripting (XSS) vulnerability has been identified in MobSF versions ≤ 4.3.2. The vulnerability arises from improper sanitization of user-supplied SVG files during the Android APK analysis workflow. When an Android Studio project contains a malicious SVG file as an app icon (e.g path, /app/src/main/res/mipmap-hdpi/ic_launcher.svg), and the project is zipped and uploaded to MobSF, the tool processes and extracts the contents without validating or sanitizing the SVG. Upcon ZIP extraction this icon file is saved by MobSF to: user/.MobSF/downloads/<filename>.svg This file becomes publicly accessible via the web interface at: http://127.0.0.1:8081/download/filename.svg If the SVG contains embedded JavaScript (e.g., an XSS payload), accessing this URL via a browser leads to the execution of the script in the context of th...