ICT Security-Sécurité PC et Internet

Anthropic’s Claude Code AI coding assistant harbored a critical network sandbox bypass for over five months, allowing attackers to exfiltrate credentials, source code, and environment variables from developer systems, and the company issued no public advisory for either incident.

Learn more / En savoir plus / Mehr erfahren:

https://www.scoop.it/t/21st-century-innovative-technologies-and-developments/?&tag=AI

https://www.scoop.it/topic/securite-pc-et-internet?tag=Claude

Attackers are abusing Google Ads and legitimate Claude.ai shared chats in an active malvertising campaign. Users searching for "Claude mac download" may come across sponsored search results that list claude.ai as the target website, but lead to instructions that install malware on their Mac.

Learn more / En savoir plus / Mehr erfahren:

https://www.scoop.it/t/21st-century-innovative-technologies-and-developments/?&tag=AI

These include a tools system to facilitate various capabilities like file read or bash execution, a query engine to handle LLM API calls and orchestration, multi-agent orchestration to spawn "sub-agents" or swarms to carry out complex tasks, and a bidirectional communication layer that connects IDE extensions to Claude Code CLI.

The leak has also shed light on a feature called KAIROS that allows Claude Code to operate as a persistent, background agent that can periodically fix errors or run tasks on its own without waiting for human input, and even send push notifications to users. Complementing this proactive mode is a new "dream" mode that will allow Claude to constantly think in the background to develop ideas and iterate existing ones.

Learn more / En savoir plus / Mehr erfahren:

https://www.scoop.it/t/21st-century-innovative-technologies-and-developments/?&tag=AI

Researchers jailbreak AI chatbots, including ChatGPT
Like a magic wand that turns chatbots evil.

A hacker was quick to pounce on the accidental leak of Anthropic’s AI tool, Claude Code, by spreading malware on a GitHub page that claimed to host the source code. 

Cybersecurity vendor Zscaler spotted a hacker exploiting interest in the Claude Code leak to push two malware strains, Vidar and Ghostsocks. Zscaler traced the threat to a GitHub page from the account “idbzoomh,” which purports to offer the leaked source code for Claude Code and claims: “I spent significant effort rebuilding the entire build system from scratch, fixing every compilation error, and making this source snapshot actually work.

Learn more / En savoir plus / Mehr erfahren:

https://www.scoop.it/t/21st-century-innovative-technologies-and-developments/?&tag=AI