#git

### Duplicate Advisory This advisory has been withdrawn because it is a duplicate of GHSA-36rr-ww3j-vrjv. This link is maintained to preserve external references. ### Original Description The Keras Model.load_model method can be exploited to achieve arbitrary code execution, even with safe_mode=True. One can create a specially crafted .h5/.hdf5 model archive that, when loaded via Model.load_model, will trigger arbitrary code to be executed. This is achieved by crafting a special .h5 archive file that uses the Lambda layer feature of keras which allows arbitrary Python code in the form of pickled code. The vulnerability comes from the fact that the safe_mode=True option is not honored when reading .h5 archives. Note that the .h5/.hdf5 format is a legacy format supported by Keras 3 for backwards compatibility.

Cybersecurity researchers have discerned evidence of two Russian hacking groups Gamaredon and Turla collaborating together to target and co-comprise Ukrainian entities. Slovak cybersecurity company ESET said it observed the Gamaredon tools PteroGraphin and PteroOdd being used to execute Turla group's Kazuar backdoor on an endpoint in Ukraine in February 2025, indicating that Turla is very likely

Law enforcement authorities in the U.K. have arrested two teen members of the Scattered Spider hacking group in connection with their alleged participation in an August 2024 cyber attack targeting Transport for London (TfL), the city's public transportation agency. Thalha Jubair (aka EarthtoStar, Brad, Austin, and @autistic), 19, from East London and Owen Flowers, 18, from Walsall, West Midlands

In the @digitalocean/do-markdownit package through 1.16.1 (in npm), the callout and fence_environment plugins perform .includes substring matching if allowedClasses or allowedEnvironments is a string (instead of an array).

The former CIA deputy director for digital innovation discusses resilience, cultural shifts, and cyber fundamentals in the AI era.

### Summary We identified a cross-site scripting (XSS) vulnerability when handling chat message in lobe-chat that can be escalated to remote code execution on the user’s machine. Any party capable of injecting content into chat messages, such as hosting a malicious page for prompt injection, operating a compromised MCP server, or leveraging tool integrations, can exploit this vulnerability. ### Vulnerability Details **XSS via SVG Rendering** In lobe-chat, when the response from the server is like `<lobeArtifact identifier="ai-new-interpretation" ...>` , it will be rendered with the `lobeArtifact` node, instead of the plain text. https://github.com/lobehub/lobe-chat/blob/0a1dcf943ea294e35acbe57d07f7974efede8e2e/src/features/Conversation/components/MarkdownElements/LobeArtifact/rehypePlugin.ts#L50-L68 https://github.com/lobehub/lobe-chat/blob/0a1dcf943ea294e35acbe57d07f7974efede8e2e/src/features/Conversation/components/MarkdownElements/LobeArtifact/index.ts#L7-L11 https://github...

This edition pulls the curtain aside to show the realities of the VPN Filter campaign. Joe reflects on the struggle to prevent burnout in a world constantly on fire.

A pair of flaws in Microsoft's Entra ID identity and access management system could have allowed an attacker to gain access to virtually all Azure customer accounts.

Microsoft and Cloudflare have delivered a major blow to the fastest growing Phishing-as-a-Service operation called RaccoonO365.

Cybersecurity researchers have discovered a new malware loader codenamed CountLoader that has been put to use by Russian ransomware gangs to deliver post-exploitation tools like Cobalt Strike and AdaptixC2, and a remote access trojan known as PureHVNC RAT. "CountLoader is being used either as part of an Initial Access Broker's (IAB) toolset or by a ransomware affiliate with ties to the LockBit,